Physical structure for a tunable sensor system for particle detection

Abstract

To provide a sensor for detecting particles.SOLUTION: A sensor comprises a silicon wafer substrate and a charge detection layer mounted on the silicon wafer substrate, and the charge detection layer has a plurality of individual pixel sensors. The sensor further includes a converter material. The converter material is operable to interact with particles of a first type and generate a reaction. The reaction generates charged particles, and the charge detection layer is configured to detect the charged particles generated by the reaction. Further, the sensor comprises a substrate layer operable to screen particles of a second type. The converter material is coated on a lower surface of the substrate layer so that the converter material faces the charge detection layer and an air gap is formed between the converter material and the charge detection layer.SELECTED DRAWING: Figure 10

Description

[Technical Field]

[0001] FIELD OF THE INVENTION Embodiments according to the present invention relate generally to detecting subatomic particles, and more particularly to devices for detecting subatomic particles. [Background technology]

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS This application is a modification of and claims priority to and benefit of Provisional Patent Application No. 62 / 626,513, entitled "Subatomic Particle Detection System For Decommissioning Activities," having a filing date of February 5, 2018, which is incorporated herein by reference in its entirety.

[0003] This application is related to U.S. patent application Ser. No. 13 / 894,305, filed May 14, 2013, now issued as U.S. Patent No. 9,435,755, entitled "SCALABLE AND TUNABLE NEUTRON DETECTION INSTRUMENT," naming Anshuman Roy as the inventor, and bearing Attorney Docket No. RHBS-0005.US, which is incorporated herein by reference in its entirety for all purposes.

[0004] This application is related to U.S. patent application Ser. No. 13 / 894,272, filed May 14, 2013, now issued as U.S. Patent No. 9,435,897, entitled "Tunable Detection Instrument for Subatomic Particulates," naming Anshuman Roy as the inventor, and bearing Docket No. RHBS-0004.US, which is incorporated herein by reference in its entirety for all purposes.

[0005] This application is related to U.S. Patent Application No. 16 / 599,959, filed August 9, 2018, entitled "PHYSICAL STRUCTURE FOR A TUNABLE SENSOR SYSTEM FOR PARTICLE DETECTION," naming Anshuman Roy as the inventor, and bearing Docket No. RHBS-0007.US1, which is incorporated herein by reference in its entirety for all purposes.

[0006] This application is related to U.S. Patent Application No. 16 / 100024, filed August 9, 2018, entitled "COMPONENT CONFIGURATION FOR A ROBUST TUNABLE SENSOR SYSTEM FOR A HIGH RADIATION ENVIRONMENT," naming Anshuman Roy as the inventor, and bearing Docket No. RHBS-0007.US2, which is incorporated herein by reference in its entirety for all purposes.

[0007] This application is related to U.S. Patent Application No. 16 / 100043, filed August 9, 2018, entitled "METHOD AND APPARATUS FOR PERFORMING PATTERN RECOGNITION FOR A TUNABLE SENSOR SYSTEM TO DETECT NEUTRON AND GAMMA PARTICLES," which lists Anshuman Roy as the inventor and has Docket No. RHBS-0007.US3, and which is incorporated herein by reference in its entirety for all purposes.

[0008] Neutrons are subatomic particles with no net electric charge. Neutrons and another subatomic particle, the proton, together form the atomic nuclei of all elements in the periodic table except hydrogen. Free neutrons are produced as a result of either nuclear fission, the radioactive decay of elements, or nuclear fusion. Special nuclear materials ("SNM"), such as plutonium used to create dirty bombs, decay radioactively to produce neutrons. Detecting such neutrons is an effective way to track the source of SNM. However, because neutrons do not carry any electric charge, their detection is problematic compared to other charged subatomic particles. One method of neutron detection that has been successfully used is to use materials that can capture incident neutrons and convert them into other easily detectable subatomic particles, such as alpha particles, tritons, and gamma rays.

[0009] Historically, high-pressure helium-3 (He3) tubes have been the mainstay of neutron detection. Neutrons striking these tubes interact with He3 nuclei to produce tritons and protium, both of which are highly energetic, charged subatomic particles that migrate toward electrodes in the presence of a strong electric field within the tube. Unfortunately, global He3 supplies are scarce, and the price of He3 has risen 20-fold in recent years in the last decade alone. Thus, there is strong consensus in the art to replace He3 technology with alternative technologies, most commonly scintillation-based detection systems and boron-coated proportional tubes.

[0010] Scintillation detectors also have several limitations. First, scintillation crystals are expensive and, due to a limited market, are produced in small quantities. Second, complex pulse shape discrimination algorithms must be utilized in these systems to distinguish neutrons from gamma rays, which also interact violently with the scintillation crystals. Due to the use of scintillating crystals, which can be sensitive to environmental factors such as humidity and salinity, scintillation detectors also suffer from reliability issues in the field. The gamma discrimination capabilities of boron-coated tubes are better than those of scintillator detectors. However, as a proportional counter technology, boron-coated tubes are limited by their form factor in the range of their applications. Furthermore, there is no global supply chain to drive down their cost over time. Both scintillation- and proportional counter-based systems must contend with significant system-level noise that interferes with measuring low incident neutron flux levels close to the cosmic background radiation. They also lack modularity, flexibility to detect subatomic particles other than neutrons, and the potential for rapid scalability.

[0011] Furthermore, conventional methods of particle detection are generally too sensitive to extreme environmental conditions. For example, if a conventional particle detection system were used to detect neutrons in a nuclear power plant or reactor after an accident, it would likely destroy any known system for radiation detection. Summary of the Invention [Means for solving the problem]

[0012] Therefore, what is needed is a technology for neutron detection that utilizes readily available and easily replaceable components that are designed to be easily adjustable and modular to detect neutrons. Furthermore, the technology needs to be flexible so that other subatomic or other particles besides neutrons can also be detected. Furthermore, the technology needs to be robust enough to detect emitted particles in intense radiation environments.

[0013] Disclosed herein is a modular, tunable technology platform with simple, readily available, off-the-shelf components that can be modified and assembled together to form highly sensitive, high-performance instruments. The off-the-shelf components used to assemble the device can be tuned to be sensitive to different particles, including neutrons. The readily available and easily replaceable components of the invention can be tuned to be sensitive to neutrons of different energies. The architecture of the disclosed embodiments of the invention enables rapid, sensitive, and flexible detection and imaging of neutrons, particularly thermal neutrons, as well as a wide variety of other subatomic particles that may accompany neutrons emitted from SNMs or other radioactive sources. The system architecture also enables identification of the element (radionuclide) that serves as the source of the incident neutrons. The architecture also enables tracking the direction of the neutron source and identification of the radionuclide or non-radionuclide source from which the neutrons originated. Finally, the architecture of the embodiments of the present invention disclosed herein allows for real-time gamma discrimination, thereby reducing false positives and instrument response times.

[0014] In one embodiment, a method for detecting particles is disclosed. The method includes generating a response to a plurality of particles using a converter material, the converter material operable to interact with the plurality of particles, a subset of the plurality of particles including neutrons. The method further includes converting a response to the response into a readable electrical signal using a sensor, the sensor comprising an array of pixels. The method also includes processing the readable electrical signal from the sensor to generate information for each pixel on the array of pixels and transmitting the information to a processing unit. Finally, the method includes performing a discrimination procedure using the information to distinguish between instances of neutron and non-neutron particle impact on the array of pixels.

[0015] In one embodiment, an apparatus for detecting neutrons is disclosed. The apparatus includes a converter layer operable to interact with and generate responses to a plurality of particles, a subset of the plurality of particles including neutrons. It also includes a sensor coupled to the converter layer, the sensor operable to convert a response to the responses into a readable electrical signal, the sensor comprising an array of individual pixel sensors, each having a respective (x, y) coordinate within the array. The apparatus further includes a first processing device operable to process the readable electrical signal to generate information about each pixel on the array, and a second processing device communicatively coupled to the first processing device. The second processing device is configured to (a) control the first processing device, (b) receive information from the first processing device, and (c) perform a discrimination procedure using the information to distinguish between instances of neutron and non-neutron particle impact on the array of pixels.

[0016] In one embodiment, a system for detecting neutrons is disclosed. The system includes a plurality of sensor modules, each sensor module including a plurality of sensor elements and a first processing device. Each of the sensor elements includes at least one converter layer operable to interact with and generate a response to a plurality of particles, a subset of the plurality of particles including neutrons. Each sensor element also includes a sensor coupled to the at least one converter layer, the sensor operable to convert a response to the response into a readable electrical signal. Further, the sensor includes an array of individual pixel sensors, each having a respective (x, y) coordinate within the array. The system can also include a second processing device communicatively coupled to the plurality of sensor modules, the second processing device operable to read information related to the respective readable electrical signals from the respective first processing devices on each of the plurality of sensor modules. Furthermore, the second processing device is operable to perform a discrimination procedure using the information to distinguish between instances of neutron and non-neutron particle impact on respective arrays of pixel sensors associated with the plurality of sensor modules. The system further includes a housing for enclosing the plurality of sensor modules, at least one of the plurality of sensor modules being tuned to detect neutrons and at least one of the plurality of sensor modules being tuned to detect non-neutron particles.

[0017] In one embodiment, a sensor for detecting particles is disclosed. The sensor includes a silicon wafer substrate and a charge detection layer disposed on the silicon wafer substrate, the charge detection layer including a plurality of individual pixel sensors. The sensor also includes a converter material operable to interact with a first type of particle to generate a reaction, the reaction generating charged particles, the charge detection layer configured to detect the charged particles generated by the reaction, and the charge detection layer configured to generate a readable electrical signal using information about the detected charged particles. The sensor further includes a substrate layer operable to filter out a second type of particle, the converter material coated on a lower surface of the substrate layer, the converter material facing the charge detection layer, and an air gap formed between the converter material and the charge detection layer.

[0018] In another embodiment, a sensor for detecting particles is disclosed. The sensor includes a silicon wafer substrate and a charge detection layer disposed on the silicon wafer substrate, the charge detection layer including a plurality of individual pixel sensors. The sensor further includes a converter material operable to interact with a first type of particle to generate a reaction, the reaction generating charged particles, the charge detection layer configured to detect the charged particles generated by the reaction, and the charge detection layer configured to generate a readable electrical signal using information about the detected charged particles. The sensor also includes a substrate layer operable to condition a second type of particle, the interaction with the substrate layer changing a property of the second type of particle, the converter material coated on a lower surface of the substrate layer, the converter material facing the charge detection layer, and an air gap formed between the converter material and the charge detection layer.

[0019] In one embodiment, a sensor for detecting particles is disclosed, the sensor comprising: a silicon wafer substrate; and a charge detection layer disposed on the silicon wafer substrate, the charge detection layer comprising a plurality of individual pixel sensors. The sensor further comprises a converter material operable to interact with one or more types of particles to generate a reaction, the reaction producing charged particles, the charge detection layer configured to detect the charged particles produced by the reaction, and the charge detection layer configured to generate a readable electrical signal using information about the detected charged particles. The sensor also comprises a substrate layer operable to screen out particle types different from the one or more types of particles interacting with the converter material, the substrate layer being adjacent to the converter material and opposite the charge detection layer.

[0020] In one embodiment, a method of capturing and analyzing information for a particle detection system is disclosed. The method includes generating a response to a plurality of particles using a converter material, the converter material operable to interact with the plurality of particles. The method further includes converting a response to the response to the reaction into an electrical signal using a plurality of sensors, the converter material operable to be coated on the plurality of sensors, each of the plurality of sensors comprising an array of individual pixel sensors, each having a respective (x, y) coordinate within the array. The method further includes processing the electrical signals to generate data for each pixel on the array of individual pixel sensors, and serializing the data collected from the plurality of sensors and transmitting the data over a thin cable to a processing unit, the processing unit being located at a remote location away from the plurality of sensors. Finally, the method includes converting the data into a series of images comprising a visual representation of the plurality of particles impinging on the plurality of sensors.

[0021] In one embodiment, an apparatus for capturing and analyzing information for a particle detection system is disclosed. The apparatus includes a converter layer coated on a substrate and operable to interact with and generate responses to a plurality of particles, the plurality of particles including neutrons. The apparatus also includes a plurality of sensors adjacent to and facing the converter layer, the plurality of sensors operable to convert responses to the reactions into electrical signals, the sensors comprising an array of individual pixel sensors. The apparatus further includes a first processing device operable to process the electrical signals to generate information for each pixel on the array of individual pixel sensors, a data serializer for serializing the generated information, and a transmission line cable for transmitting the information to a second processing unit, the second processing unit being located at a remote location away from the plurality of sensors. The second processing device is communicatively coupled to the first processing device and configured to a) control the first processing device, b) receive information from the first processing device, and c) convert the information into a sequence of images including visual representations of the plurality of particles impinging on the plurality of sensors.

[0022] In an embodiment, a system for detecting neutrons is disclosed. The system includes a plurality of sensor arrays, each sensor array including a plurality of sensors, each sensor including: (a) a converter layer disposed on the sensor, the converter layer operable to interact with and generate a response to a plurality of particles, the plurality of particles including neutrons; (b) an array of individual pixel sensors, each having a respective (x, y) coordinate within the array, the individual pixel sensors operable to convert a response to the response into a readable electrical signal; (c) a first processing device operable to process the readable electrical signal to generate information for each pixel on the array of individual pixel sensors; and (d) a data serializer for serializing the information. The system also includes a plurality of second processing devices communicatively coupled to the plurality of sensors, each second processing device associated with a respective one of the plurality of sensors, each second processing device operable to receive the serialized information from its associated sensor using a thin cable, the plurality of second processing devices being disposed at remote locations away from the plurality of sensor arrays.

[0023] In one embodiment, a system for detecting neutrons is disclosed, the system comprising a plurality of sensor arrays, each sensor array comprising a plurality of sensors, each sensor comprising: a) a converter layer disposed on the sensor, the converter layer operable to interact with and generate a response to a plurality of particles, the plurality of particles including neutrons, b) an array of individual pixel sensors, each having a respective (x, y) coordinate within the array, the individual pixel sensors operable to convert a response to the response into a readable electrical signal, c) a first processing device operable to process the readable electrical signal to generate information for each pixel on the array of individual pixel sensors, and d) a data serializer for serializing the information, the first processing device and the data serializer disposed proximate to the respective sensor. The system also includes a plurality of second processing devices communicatively coupled to the plurality of sensors, each second processing device associated with a respective one of the plurality of sensors, each second processing device operable to receive serialized information from its associated sensor using a thin cable, the plurality of second processing devices disposed at remote locations away from the plurality of sensor arrays, and the plurality of second processing devices operable to detect particle types based on the serialized information received from the plurality of sensor arrays.

[0024] In another embodiment, a computer-implemented method for detecting neutrons in images from a tunable sensor system is disclosed. The method includes training a deep learning process to recognize known radiation-dependent signature patterns produced by neutrons in a test image. The method further includes dividing the input image into multiple frames and passing the multiple frames through the deep learning process to recognize neutrons in the multiple frames. The method then includes recombining the multiple frames to create the input image. For each pixel in the input image, the method includes examining pixels connected to the respective pixel to determine whether a signature pattern characteristic of neutrons is present in the input image, and using results from the examination to count the number of neutrons in the input image.

[0025] In another embodiment, a system for detecting neutrons in images from a tunable sensor system is disclosed. The system includes a memory for storing a plurality of test images, an input image, and instructions associated with a deep learning process and a process for detecting particles of interest in the images. The system also includes a processor coupled to the memory, the processor configured to operate according to the instructions to: a) train the deep learning process to recognize a known radiation-dependent signature pattern produced by the particle of interest in the test image; b) divide the input image into a plurality of frames; c) pass the plurality of frames through the deep learning process to recognize particles of interest in the plurality of frames; d) combine the plurality of frames to recover the input image; e) for each pixel in the input image, examine pixels connected to the respective pixel to determine whether a signature pattern characteristic of the particle of interest is present in the input image; and f) determine a count of particles of interest in the input image using the connected pixels.

[0026] The following detailed description, taken in conjunction with the accompanying drawings, provides a better understanding of the nature and advantages of the present invention. [Brief explanation of the drawings]

[0027] Embodiments of the present invention are illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings, in which like reference numerals refer to similar elements and in which: [Figure 1] FIG. 1 is a block diagram of an example computing system in which embodiments of the present disclosure may be implemented. [Figure 2] FIG. 2 is a block diagram of an example network architecture in which client systems and servers may be coupled to a network according to an embodiment of the present invention. [Figure 3] 1 is an exemplary block diagram of a subatomic particle detection system in accordance with one embodiment of the present invention; [Figure 4] FIG. 2 is a schematic block diagram illustrating an exemplary hardware configuration for interfacing a host machine with a sensor module, in accordance with an embodiment of the present invention. [Figure 5A] 1 is an exemplary block diagram of a sensor array of pixels in accordance with one embodiment of the present invention; [Figure 5B] FIG. 2 is an exemplary block diagram illustrating a cross-sectional view for each pixel in accordance with one embodiment of the present invention. [Figure 6A] 1A-1C illustrate two exemplary patterns produced by two different types of subatomic particles as detected by a pixel array of a sensor in accordance with one embodiment of the present invention. [Figure 6B] 1A-1C illustrate two exemplary patterns produced by two different types of subatomic particles as detected by a pixel array of a sensor in accordance with one embodiment of the present invention. [Figure 7] FIG. 4 illustrates an exemplary information vector generated for each pixel by an MPU, in accordance with one embodiment of the present invention. [Figure 8] 1 is a flowchart of an exemplary computer-implemented process for detecting subatomic particles, in accordance with an embodiment of the present invention. [Figure 9] FIG. 2 illustrates exemplary signatures for neutrons and gamma particles used to distinguish between the two particles, in accordance with one embodiment of the present invention. [Figure 10] FIG. 2 illustrates the physical architecture of a sensor according to an embodiment of the present invention. [Figure 11] 1 is an exemplary illustration of a cross section of a primary nuclear containment vessel (PCV) of a nuclear reactor that may contain nuclear deposits resulting from a nuclear accident. [Figure 12] 1A-1C illustrate how sensing elements can be configured in a stacked arrangement according to an embodiment of the present invention. [Figure 13] 10A-10C illustrate how sensing elements can be configured in a cubic organization, according to an embodiment of the present invention. [Figure 14] 10A-10C illustrate how a sensor cube can be made to fit within a cylindrical sensor head, according to an embodiment of the present invention. [Figure 15] 10A-10C illustrate how multiple sensor cubes can be configured to fit within a cylindrical sensor head, according to an embodiment of the present invention. [Figure 16A] 10A-10C illustrate various configurations in which sensors can be arranged to maximize sensitivity, according to an embodiment of the present invention. [Figure 16B] FIG. 10 illustrates a collimated configuration that can be used to improve directional accuracy, according to an embodiment of the present invention. [Figure 17] FIG. 2 illustrates a detector configured in the shape of a cube used to generate a debris map according to an embodiment of the present invention. [Figure 18] FIG. 10 illustrates how multiple cubic sensors can be used to enable more efficient debris mapping, in accordance with an embodiment of the present invention. [Figure 19A]FIG. 10 illustrates a cylindrical configuration that allows multiple sensors to be stacked to increase sensitivity, according to an embodiment of the present invention. [Figure 19B] FIG. 10 illustrates another type of cylindrical configuration that allows multiple sensors to be stacked for increased sensitivity, in accordance with an embodiment of the present invention. [Figure 19C] 1A-1C illustrate one type of cylindrical configuration using neutron blocks to increase directional sensitivity, in accordance with an embodiment of the present invention. [Figure 20] FIG. 1 illustrates how a CMOS device sensor and a PIN diode sensor can be combined in the same detector system, according to an embodiment of the present invention. [Figure 21A] 3 is a logic diagram illustrating how data is transmitted from sensors to command and control equipment in accordance with an embodiment of the present invention. [Figure 21B] FIG. 1 is a logic diagram illustrating how data is transmitted from a robot within a primary nuclear containment vessel (PCV) of a nuclear reactor to a safe room containing command and control equipment, in accordance with an embodiment of the present invention. [Figure 22] FIG. 10 illustrates how a sensor for a detector is separated from additional electronics, according to an embodiment of the present invention. [Figure 23A] FIG. 1 illustrates a sensor-level measurement flow diagram and how neutron and gamma counts are output from individual sensors and processed in accordance with an embodiment of the present invention. [Figure 23B] FIG. 1 is a flow diagram illustrating how sensor information is processed and output by two different types of neural networks according to an embodiment of the present invention. [Figure 24A] FIG. 10 illustrates an exemplary output of a PIN diode from which neutrons can be identified using an analog pulse neural network, in accordance with an embodiment of the present invention. [Figure 24B]FIG. 10 illustrates an exemplary output of a CMOS sensor from which neutrons can be identified using a digital pattern neural network that analyzes sensor information from the CMOS sensor, in accordance with an embodiment of the present invention. [Figure 25A] 1A-1C show representative frames from a CMOS radiation sensor in response to various levels of gamma radiation, in accordance with an embodiment of the present invention. [Figure 25B] 1A-1C show representative frames at the pixel level from a CMOS radiation sensor in response to various levels of gamma radiation, in accordance with an embodiment of the present invention. [Figure 25C] 25A and 25B are histograms of representative images from FIGS. 25A and 25B, in accordance with an embodiment of the present invention. [Figure 26A] FIG. 1 shows a collection of eight bright neutron counts with 0 Gy / hr background gamma radiation, according to an embodiment of the present invention. [Figure 26B] A close-up of counts containing at least four saturated pixels. [Figure 27A] FIG. 2 illustrates a first pixel-level image having neutron and gamma signatures in the same image according to an embodiment of the present invention. [Figure 27B] FIG. 10 illustrates a second pixel-level image with neutron and gamma signatures in the same image, according to an embodiment of the present invention. [Figure 28] FIG. 1 illustrates a pixel-level image with neutron counts under high gamma conditions, according to an embodiment of the present invention. [Figure 29] 1 is a flowchart of an exemplary computer-implemented process for detecting the presence of neutrons in an image generated from sensor information, in accordance with an embodiment of the present invention. [Figure 30] 1 is a flowchart of an exemplary computer-implemented process for analyzing an image to detect neutrons using a deep learning process, according to an embodiment of the present invention. [Figure 31]1 is a flowchart of an exemplary computer-implemented process for triangulating the source location of neutron particles, in accordance with an embodiment of the present invention. [Figure 32] 1 is a flowchart of an exemplary computer-implemented process for independently controlling sensors to ensure reliability, according to an embodiment of the present invention. [Figure 33] 1 is a flowchart of an exemplary computer-implemented process for gathering information from a tunable sensor used for particle detection in accordance with an embodiment of the present invention. [Figure 34] 1 is a flowchart of an exemplary computer-implemented process for disabling a non-functioning sensor to ensure detector reliability and extend the detector's operational life, in accordance with an embodiment of the present invention. [Figure 35] 1 is a flowchart of an exemplary computer-implemented process for conserving power and managing heat in a tunable detector system, in accordance with an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0028] Reference will now be made in detail to various embodiments of the present invention, examples of which are illustrated in the accompanying drawings. While described in conjunction with these embodiments, it will be understood that they are not intended to limit the disclosure to these embodiments. Rather, the disclosure is intended to encompass alternatives, modifications, and equivalents that may be included within the spirit and scope of the disclosure as defined by the appended claims. Furthermore, in the following detailed description of the disclosure, numerous specific details are set forth in order to provide a thorough understanding of the disclosure. However, it will be understood that the disclosure may be practiced without these specific details. In other instances, well-known methods, procedures, components, and circuits have not been described in detail as not to unnecessarily obscure aspects of the disclosure.

[0029] Some portions of the detailed descriptions which follow are presented in terms of procedures, logic blocks, processes, and other symbolic representations of operations on data bits within a computer memory. These descriptions and representations are the means used by those skilled in the data processing arts to most effectively convey the substance of their work to others skilled in the art. In this application, a procedure, logic block, process, or the like, is conceived to be a self-consistent sequence of steps or instructions leading to a desired result. The steps utilize physical manipulations of physical quantities. Usually, though not necessarily, these quantities take the form of electrical or magnetic signals capable of being stored, transferred, combined, compared, and otherwise manipulated in a computer system. It has proven convenient at times, principally for reasons of common usage, to refer to these signals as transactions, bits, values, elements, symbols, characters, samples, pixels, or the like.

[0030] It should be borne in mind, however, that all of these and similar terms are to be associated with the appropriate physical quantities and are merely convenient labels applied to these quantities. Unless specifically noted otherwise, as will be apparent from the following description, throughout this disclosure, descriptions utilizing terms such as "generate," "convert," "process," "analyze," "transmit," "allocate," "detect," "associate," "access," "erase," "release," "control," "determine," or "identify" will be understood to refer to actions and processes (e.g., flowchart 800 of FIG. 8 ) of a computer system or similar electronic computing device or processor (e.g., system 110 of FIG. 1 ). A computer system or similar electronic computing device manipulates and transforms data represented as physical (electronic) quantities in the computer system memory, registers, or other such information storage, transmission, or display device.

[0031] The embodiments described herein may be described in the general context of computer-executable instructions residing on some form of computer-readable storage medium, such as program modules, executed by one or more computers or other devices. By way of example, and not limitation, computer-readable storage media may include non-transitory computer-readable storage media and communication media, and non-transitory computer-readable storage media includes all computer-readable media except transitory propagating signals. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The functionality of the program modules may be combined or distributed as desired in various embodiments.

[0032] Computer storage media include volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer-readable instructions, data structures, program modules, or other data, including, but not limited to, random access memory (RAM), read-only memory (ROM), electrically erasable programmable ROM (EEPROM), flash memory or other memory technology, compact disc ROM (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and that can be accessed to retrieve that information.

[0033] Communication media can embody computer-executable instructions, data structures, and program modules and include any information delivery media. By way of example, and not limitation, communication media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, radio frequency (RF), infrared and other wireless media. Combinations of any of the above can also be included within the scope of computer-readable media.

[0034] 1 is a block diagram of an example computing system 110 for a neutron and other subatomic particle detection system capable of implementing embodiments of the present disclosure. Computing system 110 broadly represents any single- or multi-processor computing device or system capable of executing computer-readable instructions. Examples of computing system 110 include, without limitation, a workstation, a laptop, a client-side terminal, a server, a distributed computing system, a handheld device, or any other computing system or device. In its most basic configuration, computing system 110 may include at least one processor 114 and system memory 116.

[0035] Processor 114 generally represents any type or form of processing unit capable of processing data or interpreting and executing instructions. In one embodiment, processor 114 may receive instructions from a software application or module. These instructions may cause processor 114 to perform one or more functions of the exemplary embodiments described and / or illustrated herein.

[0036] System memory 116 generally represents any type or form of volatile or non-volatile storage device or medium capable of storing data and / or other computer-readable instructions. Examples of system memory 116 include, without limitation, RAM, ROM, flash memory, or any other suitable memory device. Although not required, in some embodiments, computing system 110 may include both volatile memory units (e.g., system memory 116) and non-volatile storage devices (e.g., primary storage device 132).

[0037] Computing system 110 may include one or more components or elements in addition to processor 114 and system memory 116. For example, in the embodiment of FIG. 1, computing system 110 includes memory controller 118, input / output (I / O) controller 120, and communication interface 122, each of which may be interconnected via communication infrastructure 112. Communication infrastructure 112 generally represents any type or form of infrastructure capable of facilitating communication between one or more components of a computing device. Examples of communication infrastructure 112 include, without limitation, communication buses (such as Industry Standard Architecture (ISA), Peripheral Component Interconnect (PCI), PCI Express (PCIe), etc.) or similar buses and networks.

[0038] Memory controller 118 generally represents any type or form of device capable of manipulating memory or data or controlling communications between one or more components of computing system 110. For example, memory controller 118 may control communications between processor 114, system memory 116, and I / O controller 120 via communications infrastructure 112.

[0039] I / O controller 120 generally represents any type or form of module capable of coordinating and / or controlling the input and output functions of a computing device. For example, I / O controller 120 may control or facilitate the transfer of data between one or more elements of computing system 110, such as processor 114, system memory 116, communications interface 122, display adapter 126, input interface 130, and storage interface 134.

[0040] Communications interface 122 broadly represents any type or form of communications device or adapter capable of facilitating communications between exemplary computing system 110 and one or more additional devices. For example, communications interface 122 may facilitate communications between computing system 110 and a private or public network including additional computing systems. Examples of communications interface 122 include, without limitation, a wired network interface (such as a network interface card), a wireless network interface (such as a wireless network interface card), a modem, and any other suitable interface. In one embodiment, communications interface 122 provides a direct connection to a remote server via a direct link to a network such as the Internet. Communications interface 122 may also provide such a connection indirectly through any other suitable connection.

[0041] Communications interface 122 may represent a host adapter configured to facilitate communication between computing system 110 and one or more additional network or storage devices over an external bus or communications channel. Examples of host adapters include, without limitation, Small Computer System Interface (SCSI) host adapters, Universal Serial Bus (USB) host adapters, IEEE (Institute of Electrical and Electronics Engineers) 1394 host adapters, Serial Advanced Technology Attachment (SATA) and enhanced SATA (eSATA) host adapters, Advanced Technology Attachment (ATA) and parallel ATA (PATA) host adapters, Fibre Channel interface adapters, or Ethernet adapters. Communications interface 122 may enable computing system 110 to participate in distributed or remote computing. For example, communications interface 122 may receive instructions from a remote device or send instructions to a remote device for execution.

[0042] 1, computing system 110 may include at least one display device 124 coupled to communications infrastructure 112 via a display adapter 126. Display device 124 generally represents any type or form of device capable of visually displaying information transferred by display adapter 126. Similarly, display adapter 126 generally represents any type or form of device configured to transfer graphics, text, and other data for display on display device 124.

[0043] 1, computing system 110 may include at least one input device 128 coupled to communications infrastructure 112 via input interface 130. Input device 128 generally represents any type or form of input device capable of providing computer- or human-generated input to computing system 110. Examples of input device 128 include, without limitation, a keyboard, a pointing device, a voice recognition device, or any other input device.

[0044] 1 , computing system 110 may include a primary storage device 132 and a backup storage device 133 coupled to communications infrastructure 112 via a storage interface 134. Storage devices 132, 133 generally represent any type or form of storage device or medium capable of storing data and / or other computer-readable instructions. For example, storage devices 132, 133 may be magnetic disk drives (e.g., so-called hard drives), floppy disk drives, magnetic tape drives, optical disk drives, flash drives, etc. Storage interface 134 generally represents any type or form of interface or device for transferring data between storage devices 132, 133 and other components of computing system 110.

[0045] In one example, database 140 may be stored within primary storage device 132. Database 140 may represent a portion of a single database or computing device, or it may represent multiple databases or computing devices. For example, database 140 may represent (stored on) a portion of computing system 110 and / or a portion of example network architecture 200 in FIG. 2 (below). Alternatively, database 140 may represent (stored on) one or more physically separate devices that can be accessed by a computing device, such as computing system 110 and / or a portion of network architecture 200.

[0046] Continuing with reference to FIG. 1 , storage devices 132, 133 may be configured to read from and / or write to removable storage units configured to store computer software, data, or other computer-readable information. Examples of suitable removable storage units include, without limitation, floppy disks, magnetic tapes, optical disks, or flash memory devices. Storage devices 132, 133 may also include other similar structures or devices for allowing computer software, data, or other computer-readable instructions to be loaded into computing system 110. For example, storage devices 132, 133 may be configured to read and / or write software, data, or other computer-readable information. Storage devices 132, 133 may be part of computing system 110 or may be separate devices accessed through other interface systems.

[0047] Many other devices or subsystems may be connected to computing system 110. Conversely, not all of the components and devices illustrated in FIG. 1 need be present to practice the embodiments described herein. The above-mentioned devices and subsystems may be interconnected in ways different from those shown in FIG. 1. Computing system 110 may utilize any number of software, firmware, and / or hardware configurations. For example, the exemplary embodiments disclosed herein may be encoded on a computer-readable medium as a computer program (also referred to as computer software, a software application, computer-readable instructions, or computer control logic).

[0048] A computer-readable medium containing a computer program may be installed in computing system 110. All or portions of the computer program stored on the computer-readable medium may then be stored in system memory 116 and / or various portions of storage devices 132, 133. When executed by processor 114, the computer program loaded into computing system 110 may cause processor 114 to perform and / or be a means for performing functions of the exemplary embodiments described and / or illustrated herein. Additionally or alternatively, the exemplary embodiments described and / or illustrated herein may be implemented in firmware and / or hardware.

[0049] A computer program for controlling the particle detection system may be stored on a computer readable medium and then stored in various portions of the system memory 116 and / or storage devices 132, 133. When executed by the processor 114, the computer program may cause the processor 114 to perform functions and / or be a means for performing functions required to perform particle detection.

[0050] 2 is a block diagram of an example network architecture 200 in which client systems 210, 220, 230 and servers 240, 245 may be coupled to a network 250. Client systems 210, 220, 230 generally represent any type or form of computing device or system, such as computing system 110 of FIG.

[0051] Similarly, servers 240, 245 generally represent computing devices or systems, such as application servers or database servers, configured to provide various database services and / or run certain software applications. Network 250 generally represents any telecommunications or computer network, including, for example, an intranet, a wide area network (WAN), a local area network (LAN), a personal area network (PAN), or the Internet.

[0052] 1 , a communications interface, such as communications interface 122, may be used to provide connectivity between each client system 210, 220, 230 and network 150. Client systems 210, 220, 230 may be able to access information on server 240 or 245 using, for example, a web browser or other client software. Such software may enable client systems 210, 220, 230 to access data hosted by server 240, server 245, storage devices 260(1)-(L), storage devices 270(1)-(N), storage devices 290(1)-(M), or intelligent storage array 295. While FIG. 2 illustrates the use of a network (such as the Internet) to exchange data, the embodiments described herein are not limited to the Internet or any particular network-based environment.

[0053] In one embodiment, all or a portion of one or more of the exemplary embodiments disclosed herein are encoded as a computer program and loaded onto and executed by server 240, server 245, storage devices 260(1)-(L), storage devices 270(1)-(N), storage devices 290(1)-(M), intelligent storage array 295, or any combination thereof. All or a portion of one or more of the exemplary embodiments disclosed herein may be encoded as a computer program and stored on server 240, executed by server 245, and distributed over network 250 to client systems 210, 220, 230.

[0054] Expandable and adjustable neutron detection instrument Embodiments of the present invention provide methods and systems for detecting neutrons and other subatomic particles. While the following description focuses primarily on subatomic particles, including neutrons, embodiments and principles of the present invention can be used to detect atomic species, e.g., ions, gases, etc., or molecular species as well.

[0055] Disclosed herein is a modular, tunable technology platform with readily available and easily acquired off-the-shelf components that are assembled together to form highly sensitive, robust, and high-performance instruments. The off-the-shelf components used to assemble the device can be tuned to be sensitive to different particles. The inventive architecture disclosed herein enables rapid, sensitive, and flexible detection and identification of a wide variety of subatomic particles, such as neutrons, gamma rays, beta particles, alpha particles, neutrinos, and muons, using the same instrument. Additionally, particle detection devices of embodiments of the present invention can be designed using solid-state electronics, which helps reduce noise and vibration.

[0056] FIG. 3 is an exemplary block diagram of a subatomic particle detection system in accordance with one embodiment of the present invention. The system may have a hierarchical architecture with elements and modules arranged in an application-specific configuration. FIG. 3 illustrates a system with "N" elements, element E1 320 through element En 325. These elements constitute the basic building blocks of the system. Each module 330 in the system may include hundreds, or even thousands, of elements. Thus, the number of elements "N" may be limited only by practical considerations.

[0057] 3, in one embodiment, each element may have a converter layer C1 390 that interacts with incident subatomic particles. In one embodiment, converter layer C1 390 can be a thin film material that can be applied directly to the sensor. The design of the system is tunable because different materials (interchangeably referred to herein as "converter materials" or "reactive materials") can be used to create C1 390 to make the elements sensitive to different subatomic particles.

[0058] For example, in one embodiment, the system 157 A mixture of several isotopes, including gadolinium, also known as "natural gadolinium" 157 Gadolinium (also known as natural boron) 10 Boron, (also known as native lithium) 6The system may be tuned for neutron detection by creating C1 with layers of reactive neutron-capturing materials, such as lithium. These converter layers may be in pure elemental or compound form, or a mixture of any combination of elements and compounds of neutron-absorbing isotopes of the elements. In a different embodiment, the system may be tuned for gamma detection by creating C1 with reactive materials that interact with gamma rays, such as cesium iodide or sodium iodide. In another embodiment, the system may be tuned for fast neutron detection by designing C1 with layers of polyethylene, paraffin wax, any compound belonging to the epoxy or silicone family, or other such hydrogen-containing materials. In one embodiment, C1 may be designed to be disposed in conjunction with hydrogen-containing materials as well.

[0059] In another embodiment, one or more groups of elements may be coated with different converter layers to make the device sensitive to multiple types of particles simultaneously. For example, in one utility application, a utility company may need a single sensor that can simultaneously detect neutrons, gamma, alpha, and beta particles. Instead of using four different types of detectors to detect these particles separately, as has been done in the past, embodiments of the present invention can have a single detector with four sensors, each detecting one of the four types of particles. Alternatively, embodiments of the present invention allow a single sensor to be segmented in multiple ways, with each segment being coated with a different converter layer, so that each segment can detect a different type of particle.

[0060] In one embodiment, C1 can be selected from the following: xenon, cadmium, hafnium, gadolinium, cobalt, samarium, lithium, titanium, europium, molybdenum, ytterbium, dysprosium, erbium, and boron in their natural or isotopically enriched forms, and compounds from the above list in their natural or isotopically enriched forms, such as, but not limited to, oxides, carbides, halides (e.g., iodides, chlorides, to name a few), and combinations of elements in mixture / alloy form, or compounds of such combinations, such as gadolinium titanate, boron carbide, dimolybdenum pentaboride (Mo2B5).

[0061] The converter layer may be deposited by vapor, liquid, or plasma deposition techniques. In one embodiment, the converter layer can be a lithium fullerene compound, C60AxLix, deposited from a solution using a solvent combination such as chlorobenzene and dichlorobenzene. In another embodiment, modified boron fullerenes can be deposited from a solution. In one embodiment, the converter layer is a (C60AxLix) fullerene that is chemically bonded or bonded to a neutron absorbing element or compound. 60 , C 70 , C 84 The fullerene molecules may also be nanotubes or graphene compounds (made from any material containing molecules belonging to the carbon-based fullerene family, such as fluorofullerenes). In this case, the carbon-based fullerene molecules can be chemically bonded to the neutron-absorbing molecules either inside (endohedral fullerenes) or outside the fullerene cage. The fullerene molecules may also be made with boron, such as boron fullerenes, in which case the carbon-based fullerene molecules are not required.

[0062] In one embodiment, each element may comprise a sensor array P1 315 of pixels that convert incident particles, such as the product of an interaction between an incident subatomic particle and C1 315, into an electrical output that may be converted from an analog to a digital signal at the pixel level, i.e., individually, through a combination of transistors and analog-to-digital converters. In one embodiment, the sensor array responds to light or charge energy generated in the coating and then detected by the sensor pixels. These transistors and analog-to-digital converters may reside in a control electronics module 310, and each element may have its own control electronics module 310. In one embodiment, the sensor array P1 315 may be an off-the-shelf sensor. For example, the sensor may be a memristor, an image sensor, a photon detector, or a photovoltaic cell, among others. The sensor may also be a type of sensor commonly used in conventional consumer electronics digital cameras.

[0063] In one embodiment, P1 315 is made from any material capable of detecting charged particles, some examples of which include semiconductor polymers, such as poly(3-hexylthiophene), poly[[9-(1-octylnonyl)-9H-carbazole-2,7-diyl]-2,5-thiophenediyl-2,1,3-benzothiadiazole-4,7-diyl-2,5-thiophenediyl], also known as PCDTBT, small organic semiconductor molecules, or inorganic semiconductors such as silicon, cadmium telluride, cadmium zinc telluride, or compound semiconductors such as gallium nitride, gallium indium arsenide, or liquid semiconductor materials, or any other material (solid, liquid, or gas) that can sense (e.g., by detecting light or charge) the products of the interaction between C1 and incident subatomic particles, including neutrons.

[0064] In one embodiment, C1 390 may comprise multiple layers of materials that interact with different subatomic particles, including neutrons or other subatomic particles such as gamma rays of different energies, or it may be a composite of various materials, each of which interacts with a different subatomic particle, or it may be a combination of the two approaches.

[0065] Furthermore, the presence of C1 390 does not preclude the possibility that incident subatomic particles, including neutrons, may interact directly with the materials comprising the sensor. For example, in one embodiment, there may be instances where the material forming the sensor pixel array is itself sensitive to incident subatomic particles, such as silicon being sensitive to gamma rays, muons, etc., or boron used for p-type doping of silicon being sensitive to neutrons. Further, by way of example, semiconductors such as silicon may be sensitive to incident subatomic particles. 157 They may be doped with materials that have a high neutron capture cross section, such as Gd, and semiconductors such as PCBM (fullerene derivative [6,6]-phenyl-C61-butyric acid methyl ester) may be chemically modified with neutron-capturing materials to make the molecule sensitive to neutrons.

[0066] In one embodiment, the particle detection system may not include converter layer C1 390 at all. Instead, the converter material that would otherwise be used to create C1 layer 390 is uniformly mixed with the sensor material used to generate pixel array P1 315. By way of example, a compound neutron capture material may be mixed with a sensor material such as a semiconducting polymer, e.g., P3HT, PCDTBT, or other small organic semiconductor molecules, or inorganic semiconductors such as silicon, CdTe, or compound semiconductors such as gallium nitride, gallium indium arsenide, or liquid semiconductor material. Furthermore, P1 315 may comprise a pixelated or uniform sensor array (or monolithic sensor array) made from a semiconductor material or a material sensitive to the product of the interaction between incident subatomic particles and a reactive material. It may also be made of a composite material that is sensitive to subatomic particles and capable of generating a readable signal.

[0067] However, dispersing the converter material within the sensor material may require printing techniques, and a special process is required to mix the converter material with the sensor. As explained above, the control electronics module 310 can be used to control the operation of the elements and to send any analog or digital signals generated by the elements to the rest of the system.

[0068] In one embodiment, each of elements E1 320 through En 325 may include a lensing device L1 305 for focusing particles toward the sensor in order to improve the sensitivity of the instrument. For example, if the particle detection system is set up to detect neutrons, the neutrons can be focused using a suitable material, such as a glass polycapillary fiber made from lead silica glass and used to focus ultra-cold to fast neutrons. Alternatively, if the particle detection system is set up to detect X-rays, the X-rays can be focused using a suitable material, such as a microstructured capillary array.

[0069] The array of elements E1 320 through E n 325, in one embodiment, is connected in a series or parallel configuration to a slave processing unit 335 (referred to herein as an “SPU”). In one embodiment, the slave processing unit 335 may comprise a field programmable gate array (“FPGA”), a complex programmable logic device (“CPLD”), a microcontroller, or the like. The slave processing unit may be located (internalized) within the elements labeled E1 through E n , thereby minimizing or entirely eliminating the need for an external processing unit such as 335. The elements, together with the SPU, form a “sensor module” 330.

[0070] One or more sensor modules may be arranged in an optimized configuration to maximize system performance. For example, multiple sensor modules can be configured to operate in parallel to increase the sensitivity of the device. Because each element may be only moderately sensitive in detecting incident particles, the overall sensitivity to detected particles can be increased by stacking two or more sensor modules 330 in parallel.

[0071] While each element may be only moderately sensitive in detecting incident particles, when several of these elements are aggregated in an appropriate architecture, these components operate in a coordinated manner to result in a highly sensitive, agile, and reliable particle detection instrument. The aggregation of sensors operating in parallel results in higher sensitivity to particles and resulting imaging compared to individual elements or modules. In one embodiment, multiple sensor modules can be mounted on a common printed circuit board and operate in parallel thereon. In a further embodiment, multiple printed circuit boards, each having at least one sensor module, can be configured to operate in parallel and detect particles to further increase the sensitivity and fidelity of the platform. In one embodiment, multiple sensor modules can all be configured to detect neutrons, making the device highly sensitive to neutrons and therefore a highly reliable neutron detection instrument.

[0072] Each of the modules can be composed of multiple elements. In one embodiment, elements E1 320 to E n 325 can be made as large or small as needed to embed them in a closed geometric area, such as inside the human body, for medical applications such as single photon emission computed tomography ("SPECT"), positron emission tomography ("PET"), etc.

[0073] In one embodiment, a subset of elements E1 320-En 325 can be configured to detect different particles than the remaining elements by coating them with a different C1 converter layer than the other elements. Thus, a single module 330 can be used to detect more than one type of subatomic particle.

[0074] Each sensor module 330 is connected wirelessly or through wires to a system-level master processing unit 345 (referred to herein as an "MPU"), which controls the operation of the SPUs on the module and processes the data it receives from the SPUs. In one embodiment, the SPU in one of the modules may be capable of functioning as an MPU. In one embodiment, the MPU may be connected to several sensor modules, each sensitive to and configured to detect a different subatomic particle. Alternatively, the MPU may be connected to several stacked sensor modules that work in concert to detect the same particle, e.g., neutrons.

[0075] In one embodiment, the MPU 345 may be part of a computing system similar to the computing system 110 of FIG. 1 described in detail above. Furthermore, like the computing system 110 in FIG. 1, the MPU 345 may include the system memory 116 and storage memories 132, 133 for storing data received from the various sensor modules. The MPU 345 may send the processed data to a display 350 having a user interface (UI) that can be used to program the entire system. The display may perform functions similar to the display device 124 described above in connection with FIG. 1.

[0076] Additionally, data from the MPUs may be wirelessly relayed through wireless module 380 to host server 370, which may perform functions similar to servers 240, 245 described with respect to FIG. 2. Each of client devices 210, 220, 230 in FIG. 2 may actually be a separate computing system with an MPU connected to its own set of SPUs and reporting the results of particle detection operations to host server 240 or 245 over network 250. For example, client devices 210, 220, 230 may be security devices installed at airports to screen passenger luggage for explosive devices. Each of the client devices may then, in turn, report the results of the screening to centrally located server 240 or 245. Results from all the various screening operations may also be stored in storage devices 260(1)-(L), storage devices 270(1)-(N), storage devices 290(1)-(M), or intelligent storage array 295. In another embodiment, the MPU 345 may relay data to the host server 370 through a wired connection (not shown) instead of through the wireless module 380 .

[0077] In one embodiment, data from the various SPUs can simply flow through the MPU and be sent to the host machine 360. The host machine, in one embodiment, can be a personal computer or tablet PC, or even a smartphone, which may be a computing system similar to computing system 110 pertaining to FIG. 1 described in detail above. The host machine in such an embodiment is connected to the MPU 345 through a communications interface similar to interface 122 described in detail above.

[0078] In this embodiment, the host machine 360 ​​is responsible for processing data received from the various SPUs instead of the MPU. However, the MPU is responsible for controlling the operation of the various SPUs connected to it. Thus, the host machine performs functions similar to computing system 110. In that case, a display 350 may be connected to the host machine 360, and a user of the system may program the system using the display connected to the host machine. Alternatively, in one embodiment, the MPU 345 may reside on the host machine 360 ​​instead of within the particle detection system enclosure 340, and the various SPUs may be controlled from within the host machine 360.

[0079] In one embodiment, the particle detection system of FIG. 3 is encapsulated by placing it in a housing 340 made from a material such as plastic or metal to protect it from elements such as temperature, humidity, and dust. The housing 340, in one embodiment, may be designed to limit the penetration of certain subatomic particles, such as photons in the visible range, ultraviolet range, or higher-energy photons such as X-rays or gamma rays. For certain applications, such as neutron detection, for example, the housing 340 may include a material such as high-density polyethylene ("HDPE") that suppresses incident neutron velocities. The design of the housing 340 and the materials used to construct it vary depending on the application for the particle detection system. For example, if the detector is used in oil and gas exploration in open ocean waters, the housing 340 needs to be constructed using materials that can withstand extremely high subsurface temperatures and pressures.

[0080] FIG. 4 is a schematic block diagram illustrating a typical hardware configuration for connecting a host machine with a sensor module. The embodiment illustrated in FIG. 4 is one in which the display 340 and UI are connected to or implemented within the host machine 360, as described above. The host machine 360 ​​is responsible for processing data it receives from the MPU 345 over the communicator bus 491. The MPU 345 is responsible for controlling the operation of the various SPUs on the sensor modules 430A-430N. The sensor modules 430A-430N perform essentially the same functions as the sensor module 330 in FIG. 3. Each of the sensor modules 430A-430N shown in FIG. 4 may be configured to detect different subatomic particles. Alternatively, as described above, the sensor modules 430A-430N may be stacked and operated in parallel to reliably detect the same particle, e.g., neutrons. By using multiple sensor modules working in concert, the sensitivity and reliability of the system can be greatly improved.

[0081] The sensor modules 430A-430N can be components of a modular array and plug directly into circuit board sockets within the particle detection chassis 450. The MPU 345 can be mounted on the same circuit board that provides the sockets for the sensor modules 430A-430N to plug into, or one of the MPUs in 430A-430N can be programmed to provide the functionality of the MPU 345, thereby eliminating the need for a separate MPU 345. Because they plug into the circuit board sockets, the sensor modules can be easily inserted into and removed from the device 450. Furthermore, the placement of the sensor modules 430A-430N can be determined based on the type of particles each sensor module is configured to detect and how sensitive the user requires the system to be to particles.

[0082] Host machine 360 ​​communicates with particle detection device 450 encapsulated within housing 340 over communicator bus 491 using communication interface 122, as illustrated in FIG. 1. Communicator bus 491 provides a high-speed electronic communication channel between host machine 360 ​​and particle detection device 450. The communicator bus may also be referred to as a backplane, a module connection enabler, or a system bus. Physically, communicator bus 491 is a high-speed, high-bandwidth duplex connection bus that may be electrical, optical, etc.

[0083] In one embodiment, the particle detection device 450 can also be used in a standalone mode, such as a handheld instrument, a backpack instrument, or the like. In this embodiment, the housing of the device 450 includes the MPU 345, the display 350, the wireless module 380, and one or more sensor modules 330, so that a user can freely use the particle detector without having to physically connect it to a host machine. In another embodiment, the particle detection device 450 can also be connected to a computing device, such as a tablet PC or smartphone, via a wired connection (e.g., Ethernet or USB) or wirelessly (e.g., Bluetooth, Wi-Fi). In this embodiment, the display 350 is not required on the detection device. As described above, the MPU 345 can be part of a computing system, similar to the computing system 110 illustrated in FIG. 1 , with associated memory and display. Such a system, along with its modules, can function as a component in a collection of systems arbitrarily placed at desired locations far apart from one another to function as agents for detecting subatomic particles over a wide geographic area, on land, underground, on water, underwater, or in any other location, including space. Data collected from the various agents may be relayed to a central host machine 370 and analyzed to prepare a map of incident particles over any geographic region.

[0084] In one embodiment, module 330 can be programmed to determine the rate of subatomic particles incident on it. Alternatively, the MPU can be programmed to collect information from its connected SPU and determine the rates of various subatomic particles incident on the entire device 450. In another embodiment, particle detection device 450 can be configured to establish the direction of incident particles by arranging modules 430A-430N in an appropriate geometric configuration, such as around a sphere or in a stacked parallel configuration. For example, the direction of neutrons can be determined by using a neutron-absorbing collimator or neutron-absorbing aperture grid in front of the detector device to block all neutrons incident on them and only allow the passage of incident neutrons through (aligned with and passing through) a window in the grid or collimator.

[0085] In yet another embodiment, appropriate design of the materials used to create sensor pixel P1 315, such as a fully depleted deep CMOS or CCD sensor made from inorganic or organic semiconductors, allows the system to determine the energy of incident subatomic particles, including neutrons, thereby enabling spectroscopy.

[0086] Furthermore, in one embodiment, the entire system, or each module within the system, or even each element within the system, can be tuned to be sensitive to different subatomic particles. For example, module 430A can be configured to be more sensitive to gamma rays, while module 430B can be configured to be more sensitive to neutrons. Conversely, in one embodiment, modules can also be configured to be insensitive to certain subatomic particles. One way to make a module insensitive to certain subatomic particles is to coat converter layer C1 with an appropriate blocking layer that reduces sensitivity to certain particles. This chemical blocking capability is an advantageous feature of the present invention because it gives the user unique flexibility to configure the system to be sensitive to a select subset of subatomic particles of interest, while being insensitive to other particles that the user may not be interested in tracking.

[0087] In one embodiment, selecting highly pixelated elements E1 320-En 325 can significantly increase the granularity of the particle detection device. For example, the more pixels an element can have, the easier it is for the system to detect the location and direction from which a particular particle came. It also makes it easier to detect the particle's energy. FIG. 5A is an exemplary block diagram of a sensor array made of pixels, according to one embodiment of the present invention. As can be seen in FIG. 5A, the greater the number of pixels on pixel array P1 315, the finer the granularity, making it easier to accurately detect the position of particle 550.

[0088] FIG. 5B is an exemplary block diagram illustrating a cross-sectional view of each pixel on a sensor array of pixels, according to one embodiment of the present invention. As described above, the sensor may be an off-the-shelf component commonly found in conventional digital cameras. The electronics for the pixel may be mounted on a silicon substrate composed of p-type silicon 550 and n-type silicon 560. The area of ​​the pixel that collects information about incident particles, e.g., photons, is a photodiode 570. The pixel may include three transistors 580 T1, T2, and T3 used to collect information captured by the photodiode. For example, when a subatomic particle or a product of a reaction between an incident subatomic particle and the converter layer (390 in FIG. 3) impinges on the photodiode 570, a voltage (or current) 595 is induced through a combination of distortion and ionization of the electron field within the photodiode and the photoelectric effect. The energy of each incident subatomic particle, or the product of the reaction between a single incident subatomic particle and the converter layer (390 in FIG. 3), impinging on the photodiode 570 determines how much charge accumulates in the pixel. If several particles are incident during the time the sensor is in an exposure state, a proportionally larger number of pixel islands will accumulate charge. Transistors are used to collect information about the accumulated charge during the acquisition cycle and communicate this information to an A / D converter in the control electronics module 310. Each pixel may report an A / D conversion value between 0 and 1024 based on the intensity of the impingement on the pixel.

[0089] In some embodiments, C1 390 may be sensitive to more than one type of subatomic particle. For example, a material that reacts with neutrons may also react with high-energy gamma rays. In another example, a converter material may interact only with neutrons, while a sensor material may interact with numerous other subatomic particles, including gamma photons, alpha particles, fast electrons, etc. In these embodiments, a discrimination process used to distinguish between different types of particles while minimizing any false positives may be performed on the MPU 345. Each subatomic particle may be unique in terms of the intensity value they produce or the pattern in which they impinge on the pixels of pixel array P1 315. The discrimination procedure contains information about the unique "signature" of every particle and uses these to discriminate between particles to ensure that no false positives are generated.

[0090] For example, an incident neutron particle interacts with the material in C1 or the material of the sensor pixel, generating one type of electrical signal, while a gamma ray generates another type of signal, or a pattern of pixel islands where charge is generated beyond the thermally generated charge. Thus, discrimination between neutrons and non-neutrons becomes much faster and simpler than in proportional tube or scintillator detection systems, which require the collection of large amounts of statistical information to implement pulse shape discrimination algorithms for particle discrimination. The proposed system is capable of detecting single neutrons and distinguishing them from single non-neutron particles, such as gamma photons.

[0091] This ability to distinguish between single neutron and non-neutron particles is made possible by a unique digital signature for each type of particle. The term digital signature here refers to the pattern of pixel islands on which charge is accumulated by the incident particle or the product of the interaction between the incident particle and the converter layer C1. Thus, not only can neutrons be distinguished from other non-neutron particles, but non-neutron particles can also be further distinguished as gamma photons, x-ray photons, alpha particles, fast electrons, etc. Furthermore, every radioactive material (or radionuclide) emits a unique family of subatomic particles. For example, highly enriched uranium emits neutrons and gamma rays. Since at least some or all of these subatomic particles are detected and differentiated in the proposed system with the aid of its discrimination procedure, the source of these particles (e.g., radionuclides / isotopes) can be identified by referencing a library of digital signatures or patterns in the system's memory or in a memory external to the system.

[0092] One application of the novel discrimination procedure of the present invention is in the oil and gas exploration industry. For example, a drill used for oil exploration can be equipped with both a source of neutrons and a particle detection system of the present invention. In addition, the drill can also be equipped with a source of gamma radiation. Both the gamma and neutron data collected with the aid of the novel discrimination procedure provide essential information about the porosity and lithology of rock formations.

[0093] Another application of the novel discrimination procedure is in the homeland security industry. For example, airport security scanners may utilize the particle detection system of the present invention to detect SNMs. However, because some individuals carry radioactivity within their bodies, they may emit high-energy gamma rays that would result in false alarms being generated by the scanner without the particle discrimination procedure of the present invention. As explained above, certain materials selected for C1 may react with both neutrons and high-energy gamma rays. Using unique "digital" signatures for neutrons, gamma rays, and other particles, the discrimination procedure of the present invention prevents the occurrence of false positives. The unique digital signatures may enable the identification and counting of gamma photons and the identification of the source from which the neutrons and other particles originate. One example of this capability is the discrimination procedure's ability to distinguish between weapons-grade plutonium sources and non-neutron (and primarily gamma-emitting) sources such as Cs, Co, or Ba.

[0094] 6A and 6B illustrate two example patterns produced by two different types of subatomic particles as detected by a pixel array of a sensor according to one embodiment of the present invention. FIG. 6A illustrates a pattern produced by virtual particle A, while FIG. 6B illustrates a pattern produced by virtual particle B. If both particle A and particle B are detected by the same sensor P1 315 because converter layer C1 390 reacts with both types of particles, or because converter layer C1 390 interacts with particle A and the sensor material interacts with particle B, a discrimination procedure is required that can distinguish between the particles without generating false positives. The discrimination procedure is programmed to recognize that particle A creates a pattern of islands of pixels of statistically different intensity from particle B, and further, that the pattern consists of clumped pixels, as opposed to the diagonal or other type of pattern produced by particle B. Thus, the discrimination procedure can use the respective signatures of particle A and particle B to distinguish them from each other.

[0095] In one embodiment, the discrimination procedure can compare the pattern created by the particle to library patterns stored in the memory 116 of the host machine 360 ​​to identify the particle and identify which of the patterns in the memory the particle most closely resembles.

[0096] It is important to note that in one embodiment of the present invention, the signature patterns of a variety of different particles can be simultaneously identified. For example, the discrimination procedure is configured to simultaneously identify both particle A and particle B in the example illustrated in Figures 6A and 6B. Furthermore, if other particles are detected in the system, they can be simultaneously identified using digital signatures as well.

[0097] 9 illustrates exemplary signatures for neutrons and gamma particles used to distinguish between the two particles in accordance with one embodiment of the present invention. The discrimination procedure described above can be configured to detect pattern 910 associated with neutron-generated alpha particles and distinguish pattern 910 from pattern 920 associated with gamma photons. Thus, the discrimination procedure can identify both neutrons and gamma photons and distinguish them from one another.

[0098] In one embodiment, a digital signature can be generated using several stacked sensor modules, such as 430A-430N in FIG. 4 . In this embodiment, the generated digital signature can be a vector in three-dimensional space. The discrimination procedure uses information, such as sensor coordinates, impact intensity based on A / D readouts from pixels, and impact time, to determine the impact pattern in three-dimensional space, compare the pattern to digital signatures stored in memory, and perform statistical matching to determine the particle's identity. Performing digital signature comparison in three-dimensional space allows for increased reliability in the system. For example, a particle with higher energy may leave a higher intensity pattern on the surface module but not on modules deeper in the stack. Or, for example, particles with higher momentum, such as neutrons and gamma photons, may leave a higher intensity pattern on modules deeper in the stack but only a lower intensity signature on the surface module. Thus, analyzing the strength of a particle's response using sensors at different layers of the module stack allows for increased fidelity and accuracy.

[0099] Furthermore, in one embodiment of the present invention, noise generated and accumulated in pixels due to heat or any other cause, particularly during the time interval of exposure, can be eliminated. Methods for reducing or completely eliminating such noise may include (a) timely resetting of individual pixels or groups of pixels, or entire rows and columns of pixels, (b) optimizing the exposure, readout, and reset time cycles so that the pixels are reset as frequently as required, and (c) modifying the temperature of operation of the sensor, such as by cooling them.

[0100] In one embodiment, the discrimination procedure may use patterns produced by charge buildup within the pixels of the underlying sensor. For example, neutrons may interact with the converter layer and undergo nuclear reactions. For example, if the converter material: 157 When gadolinium is involved, the reaction is 157 Gd+n= 158 Gd * => 158 The reaction is: Gd + gamma + x-ray + IC e- + ACK e-. In one embodiment, the products of this reaction enter the sensor and generate a buildup of charge at the pixel where they first interact. The high energy of these reaction products causes secondary ionization within the pixel, resulting in enhanced charge buildup within the pixel. Furthermore, the high kinetic energy of these reaction products also causes them to diffuse onto adjacent pixels, leaving tracks of built-up charge within the sensor. A discrimination procedure within the instrument examines these tracks and determines the particle morphology. Thus, if the discrimination procedure within the processing unit of the instrument (or module) determines that gamma rays and / or x-rays and / or IC electrons and / or ACK electrons have been found within the sensor, it is concluded that a neutron has interacted with the converter layer, and the neutron count tracked by the MPU is incremented by one.

[0101] Similarly, as another example, the converter layer may be 10 It contains boron, and the incident neutrons are 10 When interacting with boron, the following reaction occurs: 10 B+n-> 7 Li+Alpha follows. These reaction products travel roughly in opposite directions, and one of them interacts with the sensor, thereby leaving a specific buildup of charge. Alpha particles, for example, generally have a very high rate of loss of energy in semiconductors and solids. As a result, the charge buildup within the pixel is found to be specifically concentrated in only a few pixels. A discrimination procedure in the processing unit distinguishes between alpha particles (or 7Li) and can distinguish this signature from any other radiation that may be incident on the instrument, such as gamma rays. As a result, the instrument can distinguish neutrons from any other subatomic particles.

[0102] The present invention is highly scalable because not only does it use relatively cost-effective, off-the-shelf components that may be chemically tailored using an appropriate converter layer, or converter material mixed with the sensor material, but users also have the ability to incorporate as many sensor modules into a device as needed. Furthermore, because the components of the present invention are readily available and low-cost, they are relatively easy to replace. Thus, if a sensor module is damaged, it is generally less troublesome to replace it than to repair it.

[0103] Yet another advantage of the present invention is that the particle detection system's housing 340 is flexible and can be configured in a manner specifically customized for several different applications. For example, the housing may be selected in such a way that pixel arrays are stacked or tiled along the walls of a cargo container and used to detect radiation within the container being shipped. Particularly in the case of neutron detection, for example, great flexibility exists in how the pixel arrays are arranged because, with very limited exceptions, neutrons can penetrate most materials until they come into contact with the material they interact with. Also, as described above, the particle detection device 450 can be configured to establish the direction of an incident particle by arranging the modules 430A-430N therein in an appropriate geometric configuration, such as around a sphere. In this case, the housing 340 is spherical. Alternatively, in other embodiments, the system can be designed to fit into a handheld or backpack device.

[0104] In yet other embodiments, module 330 and any other printed circuit boards ("PCBs") within housing 340 may be constructed using flexible materials, allowing the system to be embedded in clothing and other areas where using rigid materials is impractical. Additionally, using flexible materials allows the detector surface area to be increased, thereby increasing the sensitivity of the system. This advantageously allows the present invention to be utilized for a variety of different applications using the same system design.

[0105] In one embodiment, MPU 345 processes data from the various SPUs to which it is connected and performs all the calculations necessary to determine whether a particular particle has been detected. MPU 345 can use information from the pixel array of elements E1 320 through En 325 to precisely determine the coordinates of pixels that tested positive for a particle. MPU 345 may generate a vector consisting of information about each pixel, including the pixel's coordinates, and information about the element and sensor module in which it is located.

[0106] FIG. 7 illustrates an exemplary information vector generated for each pixel by the MPU in accordance with one embodiment of the present invention. The information vector 700 may include, in separate fields, information regarding the pixel number or (x, y) coordinate 710, the element or sensor number 720, and the module number 730 on which the pixel 710 and element 720 reside. The vector may also include information regarding the intensity value 740 read from the pixel and a timestamp 750. This information vector may be stored in memory for further analysis, or passed to the display 350 for a user to visually analyze the data, or passed through a wired or wireless connection to a computing device (such as a tablet PC or smartphone) coupled to the detector box. Alternatively, the information may be relayed to a remote location through the wireless module 380. The MPU 345 may compare the vector received from the pixel array to various signatures of different subatomic particles stored in memory to determine or confirm the particle's identity.

[0107] Additionally, MPU 345 can be programmed to provide an alert to a user of the system if more than a critical threshold number of particles are detected over a period of time and over an area. For example, in one embodiment, when the detection instrument is implemented as a handheld instrument for homeland security applications, MPU 345 may be programmed to provide an alert on display 350 if more neutrons are detected per unit volume of the instrument than background.

[0108] 8 illustrates a flowchart 800 of an exemplary computer-controlled process for detecting subatomic particles, in accordance with an embodiment of the present invention. However, the present invention is not limited to the description provided by flowchart 800. Rather, other functional flows will be apparent to those skilled in the art from the teachings provided herein, and are within the scope and spirit of the present invention. While flowchart 800 is described with continued reference to the exemplary embodiments described above, the method is not limited to those embodiments.

[0109] In step 802, neutrons or other subatomic particles may be focused toward sensor E1 320 by using lensing device L1 305. Focusing the neutrons toward the sensor improves the sensitivity of the instrument, as explained above.

[0110] In step 804, a reaction is generated when a neutron or other subatomic particle contacts converter layer C1 390. The converter layer can interact with the incident neutron to generate a reaction, the result of which is then converted into a readable electrical signal by sensor array P1 315 of pixels using control electronics module 310 in step 806. As explained above, in one embodiment, converter layer C1 may include multiple layers of materials that interact with different subatomic particles, including neutrons, or it may be a composite of materials, each of which interacts with a different subatomic particle. Furthermore, in one embodiment, instead of being a separate layer, the C1 layer may be mixed with sensor array P1 315 itself.

[0111] In step 808, the SPU 335 processes signals from the various elements E1 320 through E n 325 to generate pixel data for each sensor. Each element E1 320 through E n 325 individually may have a moderate sensitivity for detecting incident subatomic particles, but the elements as a whole result in a high sensitivity level of detection.

[0112] In step 810, the pixel data is sent to the MPU 345. The MPU 345 controls the various SPUs connected to it, collects data from the SPUs, and analyzes the data to determine any neutron impacts on the sensor's pixels in step 812. In step 814, the MPU 345 performs a discrimination procedure used to distinguish between different types of particles without generating any false positives. For example, the MPU 345 may be programmed to discriminate between neutrons and other particles, such as high-energy gamma rays that may exist simultaneously with neutrons.

[0113] Physical Structure for Tunable Sensors Figure 10 illustrates the physical architecture of a sensor according to an embodiment of the present invention. Figure 10 illustrates a cross-sectional view of a typical sensor module 330 (described in connection with Figure 3). The sensor illustrated in Figure 10 can be a CMOS sensor or any charge-sensing device, such as a pin diode.

[0114] As explained above, neutrons are particles that do not have a detectable amount of charge. To detect them, a sensor must interact with them in a way that produces a detectable signal. For most neutron detectors, this signal tends to be a charge. Therefore, neutron detectors are almost always neutron-induced charged particle detectors.

[0115] Furthermore, as mentioned above, historically, detectors for neutrons have been analog sensors comprising a gas-filled tube with a neutron-sensitive coating on its interior wall, or some type of scintillator material that produces photons when neutrons interact with it. These photons are then detected by an underlying sensor, typically after photomultiplication.

[0116] Embodiments of the present invention are considered solid-state neutron detectors that do not rely on scintillation principles and therefore operate differently. Instead, they rely on converter-on-semiconductor technology. Converter-on-semiconductor technology utilizes a neutron-reactive layer that (1) absorbs neutrons, (2) initiates a nuclear reaction, and produces ionized reaction products. (3) The ionized reaction products create a path of ionized electrons and holes through the semiconductor, which (4) are extracted and measured by applying a voltage between the anode and cathode. Embodiments of the present invention are superior to prior art methods of detecting neutrons because semiconductors are highly conductive, thus allowing sensors to operate at lower applied voltages, e.g., 5 V to 25 V. Furthermore, neutron-reactive materials are denser, allowing for more efficient capture of ionizing radiation. Therefore, detectors can be made much thinner without sacrificing detection efficiency.

[0117] In an embodiment of the present invention, for example, a neutron-sensitive layer 1004, which acts as a trap for neutrons, is positioned in proximity to the charge-sensitive device 1006. While the description herein focuses on neutrons, as explained above, it should be noted that embodiments of the present invention can be used to detect other types of particles, for example, by modifying the converter material. Neutrons interacting with this layer 1004 produce charged particles, such as alpha particles and triton particles, which, depending on their energy, have a certain spatial range they can travel before losing all their energy. The rate at which they lose energy in a medium, also referred to as -dE / dx, is highly nonlinear in that these particles lose energy at a higher rate with distance as they slow down and become less energetic.

[0118] As shown in FIG. 10 , a typical sensor for detecting particles, such as neutrons, gamma, alpha, beta, etc., comprises a thin film of neutron converter layer 1004 and an underlying charge-sensitive semiconductor device 1014. As described above, the charge-sensitive device 1014 typically comprises a semiconductor-based CMOS device (such as a silicon CMOS sensor, a bulk heterojunction polymer diode, or an organic semiconductor-based CMOS sensor), a PIN diode, or a photovoltaic device. The pixels 1006 that detect charge on the charge-sensitive device 1014 are typically mounted on a silicon wafer substrate 1007. In other words, the charge detection layer 1006 comprises a plurality of sensing elements or pixels mounted on the substrate 1007. In one embodiment, the charge detection layer is between 5 and 300 microns wide. As described in connection with FIG. 3 , each sensing element E1 320 through E n 325 comprises a pixel. 10, there is an air gap 1005, for example, 10 to 200 μm, between the converter material 1004 and the charge detection layer 1006. In one embodiment, the gap 1005 may be a vacuum (and may, but is not necessarily, filled with air).

[0119] In one embodiment, the neutron converter layer 1004 is a thin film that may be coated on a substrate 1003, such as glass, plastic, silicon, or any other material that does not strongly interact with neutrons. The substrate 1003 may be some type of metal, such as carbon fiber, polyethylene, cadmium, high-density polyethylene, steel, aluminum, or cadmium. The purpose of the substrate may be to protect the underlying sensor and to screen out certain types of particles. Each type of substrate has its own characteristics. For example, a cadmium substrate blocks fast neutrons. A lead substrate blocks a significant amount of gamma particles. A plastic substrate may slow down neutron particles and can be used to slow down neutrons. The substrate may also be used for particle screening and conditioning. It should be noted that while the embodiment in FIG. 13 is optimized for detecting thermal neutrons, embodiments of the present invention may be optimized for detecting any type of particle.

[0120] In one embodiment, if the surface of the selected substrate 1003 is sufficiently smooth, another substrate thin film can be layered on top of the substrate (not shown in Figure 13) to filter or condition different types of particles than the primary substrate.

[0121] In one embodiment, the converter layer coating may be fabricated directly on the surface of the charge detection device, as previously described. In other words, the coating may be applied directly to the charge detection layer 1006 without an air gap 1005.

[0122] In one embodiment, thermal neutron sensitivity depends on the thickness of the converter layer. Thicker converter layers capture more neutrons than thinner layers. However, thicker layers can suffer from reaction product loss if they are too thick. The ideal thickness is: 10 3 to 7 microns for orthogonal front illumination of B; 6 For orthogonal front illumination of Li, it can be in the range of 25 to 35 microns. 10The B film can be anywhere between 1 and 10 microns, while 6 The Li thin film can be anywhere between 10 and 200 microns.

[0123] The converter material coating 1004 is typically in the form of a thin film. The coated substrate is placed over the charge detection device (e.g., a CMOS sensor, a PIN diode, etc.) with the coating facing the charge detection device. Typically, an air gap exists between the top surface of the thin film coating 1004 and the exposed surface of the charge sensitive device 1006. This air gap can be adjusted to change the sensitivity of the neutron detection sensor. In a different embodiment, as previously indicated, the device can also have the neutron sensitive thin film 1004 deposited directly on the charge sensitive surface of the underlying semiconductor device.

[0124] It should be noted that the neutron converter material 1004, together with the charge-sensitive semiconductor device 1006, constitute the neutron-sensitive component of the neutron sensor system of the present invention (hereinafter referred to as the "neutron sensing element"). The neutron sensing element can be of various types. As explained above, the charge-sensitive device 1014 can be, for example, a silicon CMOS sensor of the kind used in off-the-shelf digital cameras. These CMOS sensors are designed to use several small pixels that act as individual detectors of charge within the sensing element. The charge detection layer 1006 in FIG. 10 comprises pixels.

[0125] Neutrons incident on the converter layer 1004 generate charged particles and other reaction products as they interact with the material of the converter layer. These charged particles penetrate the bulk of the converter layer 1004, through any air gaps 1005, if present, and finally through any other layers of passivation coating above the CMOS device into the charge-sensitive CMOS device 1014. In size, these particles are much smaller than the pixel. Their charge interacts with the electron cloud in the pixel, resulting in the expulsion of electrons and the creation of holes in the silicon lattice. These electrons and holes are the charge carriers that are detected as a signal in the pixel. Because the reaction products from the neutron converter layer have some energy when they are generated, they gradually lose kinetic energy within the charge-sensitive device, resulting in a finite length scale over which the expelled charge carriers are localized.

[0126] As previously mentioned, the layer of reactive neutron capturing material, C1, converter material 1004, is 157 A mixture of several isotopes, including gadolinium, also known as "natural gadolinium" 157 Gadolinium (also known as natural boron) 10 Boron (also known as native lithium) 6 These converter layers may be in pure elemental or compound form, or a mixture of any combination of elements and compounds of neutron absorbing isotopes of the elements. For example, the converter layers may be: 6 Compounds of lithium, e.g. 6 Li-X, where X represents any halide or iodide, such as fluoride, chloride, carbonate, etc. Alternatively, the material can be 10 Compounds of B, e.g. 10 BX, where X represents carbide or boric acid. 157 The gadolinium may be in its native state or oxidized.

[0127] The materials can be in amorphous, semi-crystalline, or crystalline form. They can be deposited in thin film form using a variety of methods, including chemical vapor deposition (CVD) and liquid-state, solution-state, or sol-gel processing. Multiple layers of coatings using different materials sensitive to different types of incident particles can be deposited using a combination of methods.

[0128] Neutrons, e.g., neutrons 1018, interacting with the C1 layer 1004, depending on their energy, have a certain spatial range that they can travel before they lose all their energy. Alpha particles (α particles), e.g., particles 1019, and tritons ( 3 H) Generate particles, such as charged particles, such as particle 1020.

[0129] The air gap 1005 acts as an attenuator for the charged particles. In other words, as the charged particles move through this air gap, they gradually lose energy. The charge-sensitive portion of the charge detection device (e.g., the charge detection layer 1006) needs to be positioned relative to the converter layer 1004 so that the charged particles lose most of their energy inside the charge-sensitive layer 1006. Therefore, the distance d1 1015 is set to obtain a maximum signal in the charge detection device. In one embodiment, the range for d1 can be between 10 and 200 microns. Furthermore, a typical range for the charge detection layer is between 5 and 10 microns, and the depth of the silicon wafer substrate is approximately 300 microns.

[0130] In the exemplary embodiment, when a neutron 1018 enters the sensor, it first passes through the substrate 1003. Generally, the substrate is selected so that it does not block particles of interest, e.g., neutrons in this case. The substrate may be selected so that it blocks other types of particles, e.g., gamma rays, but it is generally transparent to the particles of interest. For example, a substrate made of sheet metal or lead is suitable for blocking gamma particles.

[0131] The neutron particles 1018 interact with the converter material 1004 to produce products. For example, if the converter material 6 In the case of lithium, the following exemplary reactions may occur: 1 0n+ 6 Li→ 4 2α+ 3 H In other words, the neutron 1018 interacts with the C1 layer 1004 and produces an alpha particle 1019 and a triton (or tritium) particle ( 3 The neutrons 1018 generate charged particles, such as neutrons (H) 1020. It should be noted that the charge of the generated particles allows them to be detected by the detection layer 1006. The charged particles generally scatter in opposite directions in response to the reaction between the neutrons and the converter layer. In other words, after the reaction occurs, the alpha particles may travel in the opposite direction to the triton particles. Nevertheless, the charge detection layer 1006 detects the charge from at least one of the charged particles, which in turn allows the sensor to signal the presence of the neutron 1018. It should also be noted that for detection to occur, the converter layer 1004 generally faces the charge detection layer 1006 so that the by-products of the reaction can be easily detected in the layer 1006.

[0132] When charged particles, such as alphas, tritons, etc., are generated following reaction with the converter material 1004, they generally travel very fast initially. However, they begin to lose energy at an exponential rate. Generally, while traveling, charged particles have a certain distance they can penetrate into the silicon substrate 1007. For example, if the converter material were deposited directly on the pixel, the charged particles could penetrate as deep as 50 microns into the silicon substrate 1007 before losing all of their energy. Therefore, an air gap is added to the sensor design to ensure that the charged particles accumulate most of their energy in the charge detection region 1006, which is the most sensitive part of the sensor. The air gap ensures that the charged particles lose some, but not all, of their energy before contacting the charge detection layer 1006. If the converter material 1004 were deposited directly on the charge detection device 1014, it is likely that most of the charged particles would penetrate too deep into the substrate and go undetected. In one embodiment, the gap distance d1 1015 is optimized so that the charged particles deposit the maximum amount of energy within the charge detection region 1006.

[0133] High-spatial resolution debris mapping application example Embodiments of the present invention can be used to perform identification and high-spatial-resolution mapping of nuclear fuel debris in or around a damaged nuclear reactor. For example, the overheated core of a damaged reactor may melt, causing core debris to leak from the reactor pressure vessel (RPV) containing the fuel rods. The core debris may then be relocated into the containment vessel, which is stabilized by the flow of cooling water. In such situations, it is important to use a flexible detector that can withstand extreme conditions without being destroyed and can be used to identify and generate a map of the fuel debris within the containment vessel.

[0134] 11 is an example diagram of a cross section of a nuclear reactor's primary nuclear containment vessel (PCV), which may contain nuclear deposits resulting from a nuclear accident. The primary containment vessel 1112 comprises a thick reinforced concrete floor 1140 at the bottom of the PCV. It further comprises a cooling water pool 1110, which may contain some of the nuclear debris.

[0135] Because almost everything inside a damaged reactor unit emits some amount of gamma, simply detecting gamma rays within the PCV does not help identify fuel debris or its location with any certainty. Further complicating fuel debris identification are the mixing of molten fuel, the unknown geometric constraints of the debris field, and background activation of non-fissile material. If a gamma detector within a damaged reactor unit registers a high level at a certain location, there is no certainty that this gamma is emanating from debris.

[0136] However, spontaneous fission neutrons emitted from the core debris, if accurately detected, can characterize the debris distribution. The challenge is that the neutron flux is low compared to the high-gamma background from deposited fission products (e.g., radioactive cesium-137). This requires detecting a low proportion of neutrons in a potentially high-gamma background environment. The high-energy environment is also lethal for most sensor equipment.

[0137] Embodiments of the present invention advantageously provide a sensor system that combines the ability to discriminate low neutron flux in a high-gamma background with sufficient robustness to survive and function in extreme energy environments. Embodiments of the present invention are sensitive to ionizing radiation via direct detection in the semiconductor and to neutrons via a neutron converter layer, e.g., layer 1004 in FIG. 10. Discrimination between gamma and neutron counts occurs in 2D 60 frames per second (fps) video via pixel intensity and 2D shape, and is contextually performed by proprietary machine learning software.

[0138] Embodiments of the present invention can be used for several potential applications in the context of fuel debris detection and spatial mapping. For example, the sensor system of the present invention can be used to map core debris in the PCV, RPV, and suppression chamber of a nuclear reactor. During debris removal, the detector can be used to separate fissile from non-fissile material. Additionally, the detector can be used for recriticality monitoring.

[0139] In one embodiment of the present invention, a self-propelled robot equipped with detectors (including multiple sensors) can be programmed to enter the PCV and take measurements of gamma radiation dose and neutron flux at various points above and below the metal grid 1140 of the PCV 1112. The neutron sensors of the present invention can remain operational while being exposed to a cumulative radiation dose of up to 1000 Gy in an environment of at least 1000 Gy / hr. By comparison, the gamma background radiation under normal conditions on Earth is 10 -4 Gy. Using embodiments of the present invention, the robot can generate a high spatial resolution debris map of the PCV and identify the location of potentially harmful radiation sources (using triangulation techniques described further below).

[0140] In one embodiment, the detector is installed within a self-propelled survey robot using a cylindrical case. Furthermore, the detector configuration and cabling are advantageously customizable so that detector performance can be optimized according to the size requirements of other applications, such as criticality monitoring. Furthermore, the neutron sensor is environmentally adaptable, which is advantageous for extreme environments, such as nuclear reactors. The sensor can perform accurate readings in high-humidity environments and underwater. In one embodiment, the neutron sensor is also remotely operable so that the sensor can be controlled even when placed in extreme environments. In one embodiment, a detector with multiple sensors can include a camera for imaging in the visible range using an LED system for illumination with high resolution without sacrificing service life.

[0141] Triangulation of source locations for neutron particles. In general, in many applications, it is necessary to know whether a particular location in space is emitting radiation (such as neutrons and gamma particles) due to the presence of an object that may have material undergoing nuclear fission or some other nuclear process. For example, in the debris mapping application described above, in the case of a damaged nuclear reactor, it is critical to identify all possible sources of radiation in order to determine whether evacuation of the area or decontamination of the area is necessary. Sensing elements, such as the SPUs of the present invention, can be geometrically configured in a variety of different ways to scan an area to identify the source of the radiation.

[0142] For example, sensing elements or SPUs can be arranged around a cube, rectangular prism, sphere, icosahedron, etc. Each of these configurations is reminiscent of a compound "eye" scanning in some or all directions, searching for neutrons and other subatomic particles.

[0143] FIG. 12 illustrates how sensing elements can be configured in a stacked arrangement according to an embodiment of the present invention. In one embodiment, sensing elements 1205 arranged in an array can be stacked in a stacked arrangement 1210 to improve detection efficiency. One or more sensor modules can be arranged in an optimized configuration to maximize system performance. For example, multiple sensor modules 1205 can be configured to operate in parallel to increase the sensitivity of the device. Because each element only needs to be moderately sensitive in detecting incident particles, the overall sensitivity to particles can be increased by stacking two or more sensor modules 1205 in parallel. FIG. 12 also illustrates a side view 1220 of the sensor array and a top view 1230 of the stacked sensor modules overlooking the stacked array.

[0144] In an exemplary embodiment, each of the sensor arrays 1260 can have a small form factor and can be designed to pack a large number of sensors. For example, in the array shown in FIG. 12, each sensor can be 1 cm by 1 cm in size. Each sensor array can be 4 cm high and 6 cm wide. Furthermore, to form a stack of arrays designed for a 10 cm by 10 cm scan area, the arrays can be stacked to a depth of 6 cm. However, it should be noted that the sensors, arrays, or stacks are not limited to a particular size and can be any size or shape. Stacked sensor array embodiments are most sensitive to particles emitted from a source facing the sensor array.

[0145] FIG. 13 illustrates how sensing elements can be configured in a cubic organization according to an embodiment of the present invention. As mentioned above, the chips can be arranged in an array on a circuit board. These circuit boards can be of any size. Each chip within a circuit board functions independently. As shown in FIG. 13, a sensor array can be arranged in a cubic configuration with a CMOS chip on the sensor circuit board along each face of the cube. This allows for multi-directional sensitivity. Furthermore, the sensor arrays on each side of the cube can be stacked to include multiple arrays parallel to each other in the orientation of the cube to increase sensitivity. Neutrons incident on each face are detected only on that face. The sensor chips form the building blocks for generating high-resolution 3D maps of debris, which can be used, for example, in the application related to the damaged nuclear reactor described above.

[0146] In one embodiment, a configuration such as the one shown in Figure 13 also includes a neutron blocking material that reduces "crosstalk" between sensors in the configuration. For example, elemental forms of natural or 10 B-reinforced boron, or in natural or compound form, such as carbide, nitride, or embedded in plastics (such as HDPE) 10 Neutron blocking material 1310, such as B-enhanced boron, blocks neutrons. In different embodiments having different particles of interest other than neutrons, the blocking material may be selected appropriately to block such particles.

[0147] As shown in FIG. 13, the sensor circuit board 1320 is easily placed around a cube of neutron blocking material 1310. The neutron blocking material 1310 is opaque to neutrons. With this configuration, neutrons incident on one face of the cube will be detected only on that face. This is because for a planar sensing element, detection sensitivity varies significantly with the angle relative to the radiation source. It is greatest when the sensing element faces the radiation source. When the sensing element is at 90 degrees to the source, sensitivity is significantly lower. For example, as shown in FIG. 13, the side 1370 of the cube facing the neutron source 1350 is most sensitive to neutrons and is likely to detect the most neutrons.

[0148] Neutrons that are not detected on one face penetrate the neutron blocking material 1310, where they are absorbed. In other words, these "undetected" neutrons never reach the circuit boards located on any of the other faces of the cube. This reduces crosstalk between different sensor arrays and ensures that a clear determination can be made regarding the direction of the radiation source. Furthermore, it allows for a cleaner mapping of the particles that can be generated for display to the user.

[0149] As mentioned above, neutrons incident on one face of the cube are mostly detected by the sensing elements on that face. Because the sensing elements can be pixelated, with each pixel serving as a detection element, the angle between the sensing element and the source also creates a gradient of detection within the pixel. Within the same sensing element (e.g., the same SPU), pixels closest to the source are more likely to detect more neutrons than pixels farther away. By using the counts and profiles of neutrons detected on each side of the cube, it is possible to determine the location of the source by triangulating the results from all elements on different sides of the cube. As shown in FIG. 13, when information from multiple sensors is processed, a 3D spatial map of neutrons generally indicates that, based on the map, most neutrons were detected on side 1370 of the cube sensor. Therefore, it can be inferred that the neutron source 1350 is directly across from side 1370 of the cube. Further analysis can be performed down to the sensor level. For example, one sensor on side 1370 may detect more neutrons than another sensor on side 1370. Using the readings from each of the sensors on each side of the cube independently allows engineers analyzing the data from the sensors to further triangulate the source of the radiation.

[0150] A neutron blocker 1310 in the core of this cube helps block neutrons incident on one side of the cube from reaching the other side of the cube, allowing for better angular and spatial resolution.

[0151] In some embodiments, neutron direction information may be obtained from the neutron itself, for example, if the neutron is a fast neutron, such as fast neutron 1345. In other words, the type of neutron detected may provide some directional information.

[0152] Some radioactive sources, such as plutonium-239, emit the full spectrum of neutrons. For example, plutonium-239 can emit slow neutrons, thermal neutrons, moderated neutrons, and fast neutrons.

[0153] Thermal neutrons 1340 or hot neutrons are generally slow moving, for example, traveling at less than 2.5 km / sec. Thermal neutrons travel slowly in a cloud-like formation with other thermal neutrons, do not contain any directional information, and are generally detected by converter material. To detect thermal neutrons, a count of thermal neutrons detected on each side of the cube (based on reaction with the converter material on each sensor) is used, and the location of the neutron source is determined by triangulating the results from all the sensing elements on different sides of the cube, as described above.

[0154] However, detecting fast neutrons, which travel at speeds exceeding 10,000 km / s, is more complicated. Fast neutrons offer an additional advantage over thermal neutrons in that, because of their speed, they also carry directional information. This directional information can be used, for example, to determine the source of the neutrons. However, fast neutrons generally cannot be detected by converter materials, which makes their detection more difficult.

[0155] In one embodiment, fast neutrons, e.g., fast neutron 1345, can interact directly with the CMOS sensor itself rather than with the coating 1004. In other words, the fast neutrons interact directly with the charge detection layer 1006. When interacting directly with the silicon that makes up the charge detection layer, the fast neutrons leave a signature on the silicon that allows a discrimination procedure to determine the direction of the fast neutrons. Fast neutrons generally interact with silicon by destroying one or more silicon atoms that they contact. Furthermore, when fast neutrons contact silicon, they release highly charged particles that leave a charge signature on the silicon from which directional information can be inferred. The directional information can then be used to ascertain the source of the neutrons.

[0156] In one embodiment, the substrate 1003 can be designed using plastic or a similar material to detect fast neutrons in the same way as thermal neutrons. The plastic contains hydrogen, which slows down the fast neutrons and turns them into thermal neutrons. The converter material 1004 can then detect the fast neutrons, which have lost all their energy after passing through the plastic substrate, in the same way as thermal neutrons.

[0157] In one embodiment, if there is a high percentage of fast neutrons in the environment, the converter material 1004 for one or more sensors in the detector can be designed to use cadmium, which has the unusual property of being able to interact with fast neutrons.

[0158] FIG. 14 illustrates how a sensor cube can be made to fit within a cylindrical sensor head, according to an embodiment of the present invention. As mentioned above, a cylindrical case can be used to install a detector within a self-propelled survey robot used to survey damaged nuclear reactors. It should be noted that the sensor head may also be a different shape other than cylindrical. In this embodiment, a cylindrical case 1410 can house at least one cube sensor 1430. As shown in FIG. 14, the sensor is arranged in a cube with a neutron blocker 1440 in the center of the cube, which allows each face of the cube to be directionally sensitive.

[0159] FIG. 15 illustrates how multiple sensor cubes can be configured to fit within a cylindrical sensor head, according to an embodiment of the present invention. The CMOS chip cube 1520 forms the basis for 3D mapping of debris. Stacked cubes 1530 can also be used for added sensitivity. In one embodiment, neutron count measurements can be made on each chip of each circuit board on each face of the cube. Because the cylindrical sensor head 1510 allows both roll and pitch motion, it can rotate the equivalent of 360 degrees, allowing the detector to effectively sense a volume of space multiple times. Each time the sensor head is moved, the inner cube 1520 is oriented slightly differently, providing new information in the form of point cloud data. The measured neutrons are then reconstructed using a tomography-like technique to generate an accurate, high-resolution 3D map of the debris.

[0160] In a typical embodiment, the cube sensor configuration allows a debris scan of a 10 cm x 10 cm area to effectively generate a map with low spatial uncertainty. In situations where the detector can get close to the debris, a spatial resolution of better than 10 cm can be achieved. In one embodiment, by making the cube 1520 smaller than 10 cm, a 3D map with a spatial resolution of less than 10 cm can also be generated.

[0161] The CMOS chips used in embodiments of the present invention offer great flexibility in the range of design and configuration options. For example, circuit boards can be easily stacked to improve sensitivity along each face of the cube. For example, stacked sensor 1530 illustrates a cube sensor in which multiple circuit boards are stacked along each face of the cube to improve sensitivity.

[0162] In one embodiment, collimation can be included with a set of stacked circuit boards or a cube configuration to further improve directional accuracy. The chips can even be arranged around a sphere to create a "compound eye."

[0163] 16A illustrates various configurations in which sensors can be arranged to maximize sensitivity, according to embodiments of the present invention. For example, a collimated configuration 1610 can be used. Alternatively, the sensor array can be arranged in a cube-shaped configuration 1650. In one embodiment, as described above, to increase sensitivity, the sensor arrays can be stacked on each face of a cube to create a stacked cube configuration 1670.

[0164] 16B illustrates a collimated configuration that can be used to improve directional accuracy, according to an embodiment of the present invention. As shown in FIG. 16B, a collimator 1690 can be used around a set of stacked sensory arrays 1680 to improve directional accuracy.

[0165] 17 illustrates a detector configured in the shape of a cube and how it is used to generate a debris map, according to an embodiment of the present invention. The cubed sensor 1710 includes a scalable design that can be used to generate debris maps with low spatial uncertainty. In one embodiment, the design can be used to effectively perform debris scans of a 10 cm x 10 cm area, although spatial resolution better than 10 cm can be achieved if the detector can get closer to the debris.

[0166] 18 illustrates how multiple cubic sensors can be used to enable more efficient debris mapping in accordance with an embodiment of the present invention. In one embodiment, several cubic sensors 1810 can be used to perform a rapid scan of the area under investigation. The ability to use multiple detector modules to perform a rapid scan advantageously allows for operational flexibility and faster results.

[0167] 19A illustrates a cylindrical configuration that allows multiple sensors to be stacked to increase sensitivity, according to an embodiment of the present invention. As described above, in a nuclear reactor, for example, a robotic unit may be equipped with a cylindrical case capable of holding multiple sensor modules 1940, allowing for rapid scanning and also providing enhanced sensitivity to particles of interest. Furthermore, the cylindrical configuration leaves space 1980 within the cylinder for cabling and connections.

[0168] Figure 19B illustrates another type of cylindrical configuration that allows multiple sensors to be stacked to increase sensitivity, according to an embodiment of the present invention. In the configuration of Figure 19B, the sensors are arranged around the circumference of a cylinder (in other words, the sensors are not limited to being in a cubic or stacked configuration). The sensor shown in Figure 19B includes end caps 1956 that allow the sensor to be used in different modes, such as sideways, backwards, etc.

[0169] FIG. 19C illustrates one type of cylindrical configuration using neutron blocking to increase directional sensitivity, in accordance with an embodiment of the present invention. The neutron blocker 1955 allows the tool to be directionally sensitive, turning the detector into a "compound eye" for neutrons. In one embodiment, sensors can be stacked around the cylinder to have more layers of sensors. For example, the sensor shown in FIG. 19C in one embodiment has sensors on both sides of a printed circuit board, one side facing outward while the other side faces the neutron blocking core 1955.

[0170] FIG. 20 illustrates how CMOS device sensors and PIN diode sensors can be combined in the same detector system, according to an embodiment of the present invention.

[0171] As mentioned above, the charge sensitive device 1014 generally comprises a semiconductor-based CMOS device (such as a silicon CMOS sensor or an organic semiconductor-based CMOS sensor), a PIN diode, or a photovoltaic device.

[0172] In one embodiment where multiple sensors are used to increase sensitivity, both silicon CMOS sensors 2030 and silicon PIN diode sensors 2040 may be used in parallel. The sensors are packaged together and connected to a power supply and data processing system by cables (of customizable length). Enhanced sensitivity and accuracy of neutron detection is achieved by cross-validation between different types of sensors.

[0173] Embodiments of the present invention use multiple sensors, multiple PIN diodes 2040, and multiple CMOS sensors 2030 because sensor redundancy increases survivability and reliability, and sensor cross-validation increases measurement accuracy. Embodiments of the present invention combine silicon-based CMOS and PIN diode technologies. CMOS sensors are highly accurate digital neutron and gamma detectors. Modular CMOS components can be easily integrated into small to large detectors based on requirements. CMOS sensors are less radiation tolerant than PINs due to their smaller feature size and have a smaller active area due to their package-to-volume ratio.

[0174] On the other hand, silicon PIN diodes are more radiation tolerant than CMOS due to their larger fabricated feature sizes. PIN diodes have high timing resolution and a higher active area package-to-volume ratio. Furthermore, PIN diodes offer enhanced neutron counting efficiency due to the large sensitive area of ​​the PIN chip, especially under limited space availability. Their modular components can be easily integrated into detectors ranging from small to large, based on application requirements. However, PIN diodes tend to be less accurate than CMOS in neutron / gamma discrimination in low neutron / high gamma dose rate (and energy) backgrounds because they utilize pulse discrimination only, not pulse discrimination combined with 2D spatial recognition (as described below).

[0175] When both CMOS sensors and PIN diode sensors are used together, the combined sensor attributes overcome the disadvantages of each individual sensor type and achieve fast, accurate detection in a variety of neutron detection environments. Using multiple CMOS and PIN diode sensors to further build redundancy into the system is crucial for applications such as debris mapping, where adverse environmental conditions are almost certain to cause some degree of sensor failure. Thus, some debris mapping applications may require as many as 256 redundant sensors. Furthermore, in conditions where the signal is extremely weak, all sensors can be turned on to increase sensitivity.

[0176] Transmitting information from sensors to command and control devices In particular, in hostile environments where extreme temperatures, humidity, and radiation exist, individually or in some combination, it is often desirable to minimize or completely eliminate components of the sensor unit that may be damaged. The likelihood of component damage is generally proportional to the density of transistors within the device. For example, FPGAs, CPUs, and GPUs are more likely to be damaged in such environments than CMOS image sensors or commonly available CMOS-based data serializers. In other words, if the sensor being used is from a typical off-the-shelf camera, the additional electronics within the camera are much more likely to be damaged than the sensor itself. Therefore, when embodiments of the present invention are used in hostile environments, it is beneficial to isolate the sensor from the additional electronics using robust cabling or other methods to extend the sensor's useful life.

[0177] In one embodiment, to address the issue of electronic components being more sensitive than sensors, cabling may be used to provide a long distance between the sensitive components and relatively more robust components. Long distances may potentially facilitate, for example, being out of high-radiation environments. In this way, the sensitive components can be placed in less hostile locations, resulting in higher system reliability and longer service life. However, spatial separation of such components also requires that data from the sensors, in turn, be transmitted over the distance of separation. Spatial considerations also often dictate that data be moved over relatively thin, manageable cables. For example, in FIG. 19 , multiple sensor modules 1940 are stacked together, and each sensor may potentially require its own cable, so the cable needs to be thin and flexible enough to fit within space 1980.

[0178] In one embodiment, each of the sensors (e.g., SPUs) on the array may send their results to an MPU (e.g., MPU 345 in FIG. 3), which coordinates the results from the various sensors and sends them over a single cable. The MPU, in one embodiment, can be co-located with the sensor. As described above, the MPU can simply be another sensor or SPU programmed to control various other sensors and collect data coming from the sensors. Also co-located with the sensor is a data serializer that can withstand extremely short environments. The data serializer can serialize information from the sensor (or MPU) and transmit it over long distances to a command and control module such as that described below in connection with FIG. 21A.

[0179] FIG. 21A is a logic diagram illustrating how data is transmitted from sensors to command and control equipment in accordance with an embodiment of the present invention.

[0180] In one embodiment, where the command and control module needs to be located at a safe and potentially large distance from the sensors, the data generated in the sensors 2110A-2110D can be serialized using a data serializer operating at frequencies in the MHz to GHz range. Serializing the data allows the cable thickness to be minimized. Because there can potentially be hundreds of sensors located within the detection area, and the cables needed to transport the data must usually be designed to fit into a tight, limited geometric area, the data cables need to be thin. Data can be transmitted over the cable using, for example, low-voltage differential signaling (LVDS). Options for transmitting data over the thin cable 2198 include twisted pair cable, coaxial cable, optical fiber, etc. In one embodiment, if the geometric area is such that an extremely thin data cable is needed and the data must be moved over long distances of more than 100 meters, the data can be formatted and moved in Internet Protocol. In this way, data can be transmitted around the world and moved directly to the cloud for processing. In another embodiment, data communication can occur over WiFi or sonar channels, and cables can be eliminated entirely.

[0181] As mentioned above, the SPU 335 or MPU 345 (not shown in FIG. 21 but co-located with sensor modules 2110A-2110D) can include an FPGA or other logic device. The FPGA in the MPU, in one embodiment, can multiplex signals from the various SPUs and attach timestamps and ID numbers to the sensor information. The command and control module (in the case where a single command and control module controls multiple sensors) can then use the timestamp and ID information to determine which of the sensors is communicating. In a different embodiment, each of the SPUs can also attach timestamps and ID numbers to information from the various sensing elements and send the information to its respective command and control module.

[0182] In one embodiment, in cases where the command and control module can be co-located with the sensors (e.g., in a less extreme environment) and the data does not need to be serialized, the information from each of the sensors can be packetized by the SPU or MPU, and header information including sensor information, date, time, etc. can be attached to each packet before sending to the command and control module.

[0183] In a different embodiment, where each sensor may communicate at a different frequency, sensor information can be transmitted directly over a transmission line. The information is then combined and aggregated by the command and control module. In yet another embodiment, each sensor is connected to its own command and control module using its own associated cable, which eliminates the need to aggregate or multiplex the information.

[0184] In one embodiment, special protective shielding materials are required for the cable, especially in cases where long distances need to be covered through hostile environments. The effects of hostile environments include degradation of the cable material and noise caused by environmental factors such as temperature and radiation. For example, a lead or tungsten shield protects the cable from a strong gamma ray environment and also results in reduced noise from high-speed electrons generated by such gamma rays that impinge on the conductor portion of the cable. In high-neutron environments, the cable must be protected from neutrons by using materials such as boron, gadolinium, and lithium.

[0185] In one embodiment, due to space constraints, the thickness of each cable cannot be greater than 3 mm in diameter. In one embodiment, both fiber optic and copper cables can be used to effectively meet the system requirements.

[0186] Both the medium used to transmit the signal and the protocol used for communication must maximize transmission speed, as the information transmitted by the sensors needs to be collected and analyzed in real time.

[0187] It should be noted that while FIG. 21A is illustrated as having four sensor modules and associated command and control modules 2120A-2120D, there is no limit to the number of sensors or command modules that can operate in parallel.

[0188] Each sensor module 2110A-2110D, in one embodiment, is coupled to a respective command and control module 2120A-2120D. The command and control module controls the sensor. In one embodiment, there can be a single command and control module that controls all sensors. In a different embodiment, each command and control module can control a subset of sensors. In the embodiment shown in FIG. 13, each command and control module is associated with a respective sensor. Having separate command and control modules builds redundancy into the system so that the detector does not go completely offline if one or more of the command and control modules fail. Each command and control module includes firmware that can be programmed to control the sensor, for example, to control when the sensor is on or off, to command the sensor to transmit its data or to stop transmitting data, etc. Additionally, the command and control module can include registers for storing data coming in from the sensor.

[0189] In one embodiment, if environmental adversity is not a factor, the command and control module can be merged with the sensor. In other words, in situations where the sensor does not need to be separated from the sensitive electronics that control it, the command and control module can be co-located with the sensor.

[0190] As discussed above in connection with FIG. 3, in one embodiment, data from the various sensors (or SPUs) can simply flow through an MPU. The MPU can be programmed to control a group of SPUs or sensors. In one embodiment, an MPU (not shown in FIG. 21) controlling the sensors (2110A-2110D) can collect information from the sensors and send it to an associated command and control module. The command and control module can be completely separate from the MPU, for example, a computer or server stack connected to the MPU. Alternatively, it can have some components within the MPU (e.g., FPGA, microprocessor, etc.) and some components outside the MPU (e.g., computer or server stack). In different embodiments, the SPUs or sensors can be programmed to communicate directly with their respective command and control modules without using an MPU.

[0191] In the reactor example described above, sensors 2110A-2110D can reside in a cylindrical casing housed within a robot sent into the PCV, while the command and control module can be located a safe distance from the reactor. Thus, the distance over which cables must communicate is typically around 60 meters or more. The sensors are connected to the command and control module using thin cables, such as twisted pair cables, coaxial cables, optical fibers, etc. Thin cables are important because when several sensors are used, several cables must pass through a closed, narrow geometric area.

[0192] In one embodiment, each of the command and control modules is programmed to use sensor information including the type of particle the sensor is optimized to detect. For example, sensor 2110A may be optimized to detect neutrons, while sensor 2110B may be optimized for the detection of alpha particles.

[0193] As mentioned above, some utility-related applications require the simultaneous sensing of alpha, neutron, gamma, and beta particles. Sensors 2110A-2110D may all be located within a single detector, with each sensor within the detector optimized to detect a different type of particle. The associated command and control module is programmed with information regarding the type of particle that each sensor module is designed to detect.

[0194] If one of the sensors is compartmentalized with a different type of coating across the sensor, the corresponding command and control module must be programmed to use that information as well.

[0195] In one embodiment, each command and control module is connected to a respective computer (e.g., computers 2130A-2130D). Alternatively, command, control, and computation can all be part of the same computing module. In other words, each computer (e.g., computers 2130A-2130D) can be integrated with a respective command and control module. Furthermore, there can be a single computer rack dedicated to each sensor, with each command, control, and computation module mounted on the rack. In one embodiment, all computation modules 2130A-2130D can be part of a single master computer that controls all connected command and control modules.

[0196] In one embodiment, the command and control module is programmed to receive serialized data from a CMOS sensor in the detector system over a twisted pair or coaxial cable and convert the serialized data into a sequence of bitmap images at a rate of 60 frames per second. In other words, the command and control module performs data formatting on the information received from the CMOS sensor and converts them into bitmap images. While the description herein focuses on bitmap images, it should be noted that the command and control module may convert sensor data into images in any format, such as jpeg, raw, png, etc.

[0197] These bitmap images can then be analyzed by a discrimination procedure running on computing modules 2130A-2130D to detect the presence of various types of particles, e.g., gamma, neutrons, etc. In one embodiment, computing modules 2130A-2130D implement a particle discrimination procedure that can analyze the bitmap images to detect the presence of particles.

[0198] Signals from the command and control modules are analyzed in real time by computing modules 2130A-2130D and sent to display 2150 (similar to display 350 described in connection with FIG. 3). Information collected from the CMOS sensors is displayed as 2D 60 fps video, and proprietary machine learning software running on computing modules 2130A-2130D is used to distinguish between gamma and neutron counts.

[0199] For example, artificial intelligence algorithms discriminate neutron particle counts in a background of gamma particles. Data processing is done in real time, and accurate discrimination in high radiation environments is achieved at the software level through artificial intelligence (AI) algorithms for accurate discrimination between neutrons and gamma particles (as further described below).

[0200] In one embodiment, the command, control, and compute modules are all programmed into firmware on the GPU. Each sensor typically transmits 1 MB of information per second, or approximately 2 TB of data in 8 hours. Therefore, a high-speed GPU is required to manage the large amount of data being transmitted. As mentioned above, the data analysis and data storage electronics associated with the command, control, and compute modules are located farther away from the sensors to ensure circuit reliability and protect them from adverse conditions.

[0201] FIG. 21B is a logic diagram illustrating how data is transmitted from a robot within a primary nuclear containment vessel (PCV) of a nuclear reactor to a safe room containing command and control equipment, in accordance with an embodiment of the present invention.

[0202] As discussed above, embodiments of the present invention advantageously provide a sensor system that combines the ability to discriminate low neutron flux in a high-gamma background. Embodiments of the present invention can be used for several potential applications in the context of fuel debris detection. For example, the sensor system of the present invention can be used to map core debris in the PCV (e.g., containment vessel 2184 shown in FIG. 21B), RPV, and suppression chamber of a nuclear reactor. During debris removal, the detector can be used to separate fissile from non-fissile material. Additionally, the detector can be used for recriticality monitoring.

[0203] As shown in Figure 21B, in one embodiment of the present invention, a detector-equipped self-propelled robot 2181 (with multiple sensors) can be programmed to enter the PCV 2184 and take gamma radiation dose and neutron flux measurements at various points above and below a metal grid 2182 (similar to metal grid 1140 shown in Figure 11) in the PCV. The neutron sensors of the present invention can remain operational while being exposed to a cumulative radiation dose of up to 1000 Gy in an environment of at least 1000 Gy / hr. Using embodiments of the present invention, the robot can generate a high-spatial resolution debris map of the PCV 2184 and identify the locations of potentially harmful radiation sources, such as nuclear debris 2183.

[0204] In one embodiment, the detector is mounted within the self-propelled survey robot 2184 using a cylindrical casing. In one embodiment, the neutron sensor is also remotely operable, as in the case of the robot 2181, so that the sensor can be controlled even when placed in extreme environments, such as the containment vessel 2184.

[0205] FIG. 21B provides an example scenario in which the command and control module needs to be located far from the sensors. Data generated at the sensors or sensors (equipped within the robot 2181) can be serialized using a data serializer operating at frequencies from MHz to GHz. Serializing the data allows cable thickness to be minimized and (if the robot 2181 is equipped with several sensors) several cables to fit within a closed geometric area. As mentioned above, options for data transmission include twisted pair cable, coaxial cable, optical fiber, etc. In one embodiment, if the geometric area is such that extremely thin data cables are required and data must be transferred over long distances of more than 100 meters, the data can be formatted and transferred in Internet Protocol.

[0206] In the scenario illustrated in Figure 21B, the sensors and data extraction logic are separated from the sensitive electronics (as also described in connection with Figure 21A). Thus, the command and control module 2192, the computing module 2193, and the display 2194 are all located in a secure room 2185 a significant distance from the containment vessel. The secure room may be several hundred meters away from the containment vessel and safe for human entry. It should be noted that while Figure 21B is illustrated as having a single sensor module, a single command and control module 2193, and a single computing module 2193, there is no limit to the number of sensors or command, control, and computing modules that can operate in parallel.

[0207] FIG. 22 illustrates how sensors for a detector are isolated from additional electronics in accordance with an embodiment of the present invention. As shown in FIG. 22, the CMOS sensor 2210, PIN sensor 2220, camera 2230, and LED lighting 2240 modules are generally more durable and can withstand extreme environments. Therefore, they can be isolated from the processor 2260, computer 2270, and power supply 2280. In one embodiment, a data aggregation module, e.g., MPU 2250, can be co-located with the CMOS and PIN sensors. For example, in cases where one of the SPUs is programmed to be an MPU or data aggregator module, it can be co-located with the other sensor modules. In one embodiment, the data aggregation module 2250 can be a data serializer that serializes data sent to the command and control module. As mentioned above, when embodiments of the present invention are used in hostile environments, it is beneficial to isolate the sensors and camera from additional electronics using rugged cabling or other methods.

[0208] Use of multiple sensors to ensure reliability In one embodiment, each of the sensors can be independently controlled using the command and control module (and associated MPU). Individual control for each sensor allows for significant flexibility in the detection system. When the signal is weak, all sensors can be turned on. Alternatively, for example, when the detector is in close proximity to fuel debris and the signal level is high, all sensors except one can be turned off. In one embodiment, the command and control module may be programmed to use logic to determine the number of sensors that can be turned off in a high-signal environment without sacrificing the integrity of the detection process. In other words, the command and control module may have a threshold number of particles or a threshold level of signal that it needs to be able to detect before it turns off additional sensors. Similarly, if the signal level falls below a certain threshold, the command and control module may be programmed to turn on additional sensors to increase strength / sensitivity.

[0209] In one embodiment, sensing elements may be turned off or disabled because they are not operating as required or have otherwise deteriorated. One way to do this is to perform a sensing element check, which includes a signal check using a pattern generator. A signal is sent from the associated command and control module to the sensor asking it to send back a pattern. Based on the deviation of this pattern from what is expected, the command, control, and computation module can determine the extent of damage and overall sensing element performance. If the damage exceeds a certain threshold, the sensor is turned off. Because the system has multiple redundant sensors, this generally does not affect the overall performance of the detector.

[0210] In one embodiment, the pattern sent to each of the sensors to determine its health is a checkerboard pattern. In other words, every other pixel is stimulated. The pattern is then read out using a command and control module. If a significant deviation from the checkerboard pattern is received, the command and control module can determine that the sensor is not functioning optimally.

[0211] The command and control module is programmed with logic to automatically deal with sensor failures, for example, the command and control module is programmed to seamlessly take a failed sensor offline and bring up a still-operational sensor from the queue.

[0212] In one application, for example, in an extreme operating environment, it may be guaranteed that more than 50% of 1000 sensors will fail. In such a situation, the command and control module may be programmed to turn on only a certain amount of sensors at a time. It has been determined that inoperative sensors have a longer useful life than operative sensors. For example, the command and control module may turn on 100 sensors to start. In adverse conditions, the on sensors are more likely to fail. Thus, when those sensors fail, the command and control module may have the multiplexing logic turn on the next 100 sensors and deactivate the failed sensor, and may continue in this manner for the duration of the sensor's mission.

[0213] As mentioned above, pixels degrade in extreme environments. Contributing factors include temperature, radiation, humidity, etc. In one embodiment, the command and control module, or the computation module, is configured to automatically correct for pixel degradation. In other words, degraded pixels may be characterized to determine how their electrical characteristics change in response to extreme environments. This characterization then allows the command and control module (or computation module) to correct for degraded pixels in real time, which allows the sensor to continue operating despite exposure to extreme environments.

[0214] 32 and 34, further described below, provide further explanation regarding how sensors can be independently controlled to ensure reliability.

[0215] Power Savings and Thermal Management Typically, sensors operate in extreme environments. This is exacerbated by the fact that CMOS sensors heat up easily. Furthermore, CMOS sensors become noisier as temperature increases. Individual sensor control and redundancy allow users to selectively turn off selected sensors, allowing for dramatic heat dissipation. In one embodiment, sensors can be cycled on and off to prevent heat buildup. For example, sensors 2110A, 2110B can be turned on while sensors 2110C, 2110D are turned off, and vice versa. Furthermore, in one embodiment, a command and control module associated with each sensor may keep track of the sensor and its operating temperature. If the command and control module senses a temperature above an acceptable threshold, it may automatically turn the sensor off.

[0216] Additionally, individual control over sensors also enhances power savings. As mentioned earlier, in a low-particle environment, when the signal is weak, all sensors can be turned on. Alternatively, in a high-particle environment, when the signal level is high, all but one, or just a few sensors, can be turned off. This allows the system to efficiently conserve heat and energy while extending the life of the sensors.

[0217] In one embodiment, the sensors are configured to allow partial shutdown. In other words, the sensors are equipped with a mode that allows them to be partially shut down while still operating in a low-power mode. For example, in low-power mode, a CMOS sensor consumes 1 / 1000 of the CMOS's maximum operating power. Because sensors typically use a significant amount of time and power to start up from an off state, enabling partial shutdown mode allows sensors that are not needed to be partially shut down instead of being completely turned off. These sensors can then be easily brought back online without requiring them to go through a sensor initialization process.

[0218] FIG. 35, described further below, contains further description of an exemplary process for conserving power and managing heat in a tunable detector system.

[0219] Pattern Recognition and Software Extraction As mentioned earlier, in some embodiments, a discrimination process can be performed (e.g., on computing modules 2130A-2130D) to distinguish between different types of particles while minimizing any false positives. Each subatomic particle may be unique in terms of the intensity values ​​they produce or the pattern in which they impinge on the pixels of pixel array P1 315. The discrimination procedure may include information about the unique "signature" of every particle and use these to discriminate between particles to ensure that no false positives are generated. The ionizing radiation-dependent signatures are the basis of the particle discrimination algorithms that run on computing modules 2130A-2130D.

[0220] For example, neutrons are observed to deposit more energy than the gamma background (gray, black), resulting in higher amplitude (white) pulses. Embodiments of the present invention include image processing procedures (e.g., within computational modules 2130A-2130D) that use artificial intelligence digital pattern recognition software to discriminate and identify neutrons, as well as count them, using amplitude and shape discrimination libraries.

[0221] As previously mentioned, embodiments of the present invention can combine the ruggedness and short detection time of PIN diodes with the excellent particle discrimination of CMOS image sensors.

[0222] FIG. 23A illustrates a sensor-level measurement flow diagram and how neutron and gamma counts are output from individual sensors and processed in accordance with an embodiment of the present invention.

[0223] In one embodiment, sensor information from the PIN diode sensor is analyzed using an analog pulse neural network 2230, while sensor information from the CMOS sensor 2320 is analyzed using a digital pattern neural network 2340. Neutron and gamma counts are determined in real time and output from both types of neural networks and passed through an inference module 2350 (which may be programmed onto calculation modules 2130A-2130D) to determine the total neutron and gamma counts. The inference model 2350 also outputs the statistical certainty of the neutron and gamma count measurements in real time.

[0224] FIG. 23B illustrates a flow diagram illustrating how sensor information is processed and output by two different types of neural networks in accordance with an embodiment of the present invention. This embodiment includes neural networks used to distinguish gamma and neutron signals in a CMOS image 2321 and a PIN diode voltage signal 2322. In one embodiment, an artificial intelligence (AI) algorithm, such as a deep fully convolutional neural network (CNN) architecture, can be used for CMOS and PIN diode sensor data processing. The neural network is pre-trained using known and experimental data generated during sensor testing and characterization. This training process provides a trained model for fast, real-time neutron and gamma classification via pattern recognition. The PIN diode input signal 2322 and the CMOS image 2321 are deconvolved to separate the high-intensity neutron signal from the low-intensity gamma background. The PIN diode input signal 2322 is deconvolved to produce output 2324 from the analog pulse neural network 2330, while the CMOS input image 2321 is deconvolved to produce output 2323 from the digital pattern neural network 2340.

[0225] For CMOS images, the digital pattern neural network 2340 is programmed to apply an artificial intelligence procedure, such as a deep CNN, to the CMOS images 2321 (input at 60 fps from the command and control module) and identify neutrons using an AI procedure trained to recognize neutron patterns in the CMOS images. The AI ​​procedure can be programmed, for example, to output the neutron-identifying CMOS images on a display. Similarly, the analog pulse neural network 2330 is programmed to apply an artificial intelligence procedure, such as a deep CNN, to the pulses generated by the PIN diodes to determine neutron and gamma counts.

[0226] FIG. 24A illustrates an example output of a PIN diode from which neutrons can be identified using an analog pulse neural network, according to an embodiment of the present invention. Neutrons absorbed in the converter layer of the PIN diode produce ionizing radiation. A portion of the ionizing radiation (energy emissions from nuclear reactions, alpha and beta particles, and gamma photons) incident on the sensor is absorbed in the active layer of the semiconductor. Detection in the PIN diode occurs in the active pixel area of ​​the PIN diode. The PIN diode is a single photodiode.

[0227] The amplifier boosts the signal from the diode by several orders of magnitude. The amplitude- and time-dependent signal from the diode is first filtered using an analog discriminator and then fed to an analog-to-digital (ADC) converter. The PIN diode readout results in a series of time- and amplitude-dependent pulses. As shown in both images in FIG. 24A, gamma photons generally generate pulses 2440 that are wider and lower than the pulses 2420 generated by neutron particles. The pulse height is determined by the amount of energy deposited by each particle; neutrons generally deposit more energy than gamma particles. Once the amplified pulses pass through the analog discriminator and ADC, software running on the command, control, and computation module saves the raw pulse data and then processes it to ultimately distinguish pulses resulting from neutron absorption from pulses resulting from gamma absorption. The resulting pulse height information can also be used to determine neutron and gamma counts. Furthermore, once the gamma and neutron counts are sorted using the analog pulse neural network 2330, the raw data is saved and input into the inference model 2350 for accurate particle counting in real time.

[0228] In one embodiment, instead of using an analog pulse neural network to identify raw pulses from a PIN diode sensor, a digital pattern neural network may be programmed to convert the image of the pulses in Figure 24A (obtained from an oscilloscope or otherwise) into a bitmap image and train an artificial intelligence algorithm, e.g., a deep CNN, to identify and differentiate neutron-related pulses 2420 from gamma-related pulses 2440 directly from the bitmap image. For example, a command, control, and computation module associated with the PIN diode can be programmed to analyze bitmap (or jpeg, png, raw, etc.) images of neutron and gamma pulses generated by the PIN diode sensor and identify neutrons directly from the bitmap image.

[0229] 24B illustrates an exemplary output of a CMOS sensor from which neutrons can be identified using a digital pattern neural network that analyzes sensor information from the CMOS sensor, according to an embodiment of the present invention. As mentioned above, neutrons are observed to deposit more energy than the gamma background, resulting in higher amplitude (white) pulses. Image processing software running on computational modules 2130A-2130D uses amplitude and shape discrimination libraries to discriminate, identify, and count neutrons 2444 from bitmap images generated at 60 fps by the CMOS sensor using artificial intelligence digital pattern recognition software.

[0230] In one embodiment, the CMOS pixel-level data is read out and rendered into a bitmap image similar to the one shown in Figure 24B using a digital pattern neural network 2340. In high gamma background conditions, the readout rate may be increased to reduce pileup. The neutron and gamma counts are input into an inference model 2350 for accurate particle counting in real time.

[0231] As mentioned earlier, raw data from a CMOS-based radiation sensor comes in the form of a series of images at approximately 60 frames per second. An image contains 360,960 pixels, each 6.0 μm by 6.0 μm in size. In one embodiment, each pixel records data in the range of 0 to 255.

[0232] Figure 25A illustrates representative frames from a CMOS radiation sensor in response to various levels of gamma radiation, in accordance with an embodiment of the present invention. In a radiation source-free environment (0 Gy / hr), the CMOS-based radiation sensor records a background pixel response of 7 to 8. Frame 2510A therefore appears black. As the gamma dose rate increases (from 0 Gy / hr to 1200 Gy / hr in image 2520A), the overall response rises. The frame appears brighter.

[0233] Figure 25B illustrates representative frames at the pixel level from a CMOS radiation sensor in response to various levels of gamma radiation, in accordance with an embodiment of the present invention. Figure 25B illustrates the same frames as Figure 25A, but at the pixel level. Similar to Figure 25A, frame 2510B at 0 Gy / hr appears black, while frame 2510B at 1200 Gy / hr appears bright due to a higher degree of background interference.

[0234] Figure 25C illustrates histograms of representative images from Figures 25A and 25B, in accordance with an embodiment of the present invention. As can be seen in the histograms corresponding to the images in Figures 25A and 25B, the mean response increases as the dose rate increases from 7.95 at 0 Gy / hr to 90.7 at 1200 Gy / hr. It is also observed from the histograms that the standard deviation appears to change significantly at low dose rates. For example, the standard deviation of the images at 0 Gy / hr is less than 1, and at 62 Gy / hr it increases by an order of magnitude (6.711).

[0235] To train an artificial intelligence algorithm, such as a deep CNN, to recognize neutron counts from images with various levels of gamma radiation, neutron patterns first need to be established based on existing knowledge of neutron pattern symmetry and the underlying physics of neutron behavior. Furthermore, neutrons need to be labeled in the training images based on these patterns.

[0236] FIG. 26A illustrates a collection of eight bright neutron counts with 0 Gy / hr background gamma radiation, according to an embodiment of the present invention. The images in FIG. 26 may be a series of training images input into artificial intelligence software to train it to recognize neutrons from images with various levels of gamma radiation. Qualitative similarities in the patterns among the neutron counts are observable from the eight images in FIG. 26. For example, all counts contain a bright and symmetrical pattern characterized by a degree of radial symmetry, a saturated central pixel, and pixel intensity gradually decreasing from the center. Each count appears to contain at least four saturated pixels 2690, with the intensity of the saturated pixels equaling 255. For example, the count in frame 2620 contains at least six saturated pixels.

[0237] Figure 26B illustrates a close-up of a count containing at least four saturated pixels. Figure 26B illustrates the count in frame 2620 with pixel intensity values ​​overlaid. As can be seen in Figure 26B, the count in frame 2620 contains at least six saturated pixels 2690.

[0238] Counts containing saturated pixels are easy to identify by eye, and because the counts deposit large amounts of energy in a distinguishable pattern within the CMOS pixels, they are also distinguishable by software, even in the presence of significant amounts of gamma radiation.

[0239] Once neutrons are labeled based on patterns in training images, a deep learning artificial neural network, e.g., a deep CNN, can be trained to recognize these patterns. Other types of artificial intelligence algorithms that can be trained to recognize neutrons in images include, for example, ReLu CNN, cascaded CNN, support vector machines, random forests, XG Boost, various types of LSTM, recurrent neural networks, convolutional deep neural networks, and Bayesian deep neural networks. After the deep learning software is trained, it can be tested on a new set of images in which neutrons were previously identified. In other words, the deep learning software can be tested on images in which the number and location of neutrons are known. If the accuracy is acceptable, new images can be fed to the artificial neural network; otherwise, training is performed using more images. In one embodiment, if the deep learning software flags a false positive, it can be retrained to recognize the patterns that resulted in the false positive and refrain from flagging them in the future.

[0240] 27A illustrates a first pixel-level image having neutron and gamma signatures in the same image in accordance with an embodiment of the present invention. In one embodiment, a detector of the present invention detects gamma radiation and neutrons in a single frame. Neutron and gamma signatures are easily distinguished due to significant differences in their physical interactions. Neutron signatures deposit a significant amount of energy in the CMOS pixels and retain their distinct size and shape patterns even when gamma interactions increase the brightness of the surrounding image.

[0241] The image in Figure 27A contains a relatively low level of background gamma radiation, approximately 62 Gy / hr. Therefore, the neutron pattern 2702 is easily distinguishable. The image in Figure 27 is an example of neutron detection in an environment with neutron flux and gamma radiation. The image contains one bright neutron count and the clear signature 2704 of a gamma photon.

[0242] As mentioned previously, neutrons interact with the converter layer adjacent to the CMOS pixel. The reaction between the neutrons and the converter layer produces alpha particles and tritons, ionizing particles that have a finite probability of interaction with silicon. Gamma photons interact directly with silicon via Compton scattering, depositing some of their energy in the crystal lattice. The resulting change in voltage within the CMOS active pixel is measured.

[0243] It should be noted that embodiments of the present invention can also be used to detect gamma photons, for example, using a CMOS sensor. As mentioned, gamma photons interact with silicon and light up pixels as shown in Figure 27A with a distinct gamma signature 2704.

[0244] The neutron counts in the image of Figure 27A are easily identifiable because they contain a characteristic brightness and symmetric pattern with a degree of radial symmetry, a saturated central pixel, and pixel intensity that gradually decreases from the center. Furthermore, compared to the pixel intensities of neighboring gamma photons 2704, the maximum pixel intensity of the neutron counts is much higher. For example, in Figure 27A, the maximum pixel intensity accumulated by the gamma photon scattering interaction is 87, and the total pixel intensity for the six brightest pixels in the gamma track is 200 to 300. By comparison, the maximum pixel intensity for the neutron counts is 255 (saturated), and the sum of the bright pixels is greater than 3000.

[0245] In one embodiment, the command and control module can be programmed to recognize damaged pixels within the sensor. Damaged sensors can be detected because they typically remain at a high intensity value for an extended period of time. For example, a damaged pixel may provide a consistent reading of 255 over several consecutive frames. The command and control module can be programmed to selectively turn off certain pixels within the sensor. In other words, at the software level, the command and control module can be programmed to recognize that certain pixels are damaged and to ignore the output from those pixels.

[0246] FIG. 27B illustrates a second pixel-level image with neutron and gamma signatures in the same image, according to an embodiment of the present invention. As seen in FIG. 27B, neutron counts 2780 are easily identifiable because they contain a characteristic bright and symmetrical pattern with a degree of radial symmetry, a saturated central pixel, and pixel intensity gradually decreasing from the center. On the other hand, gamma patterns 2750, 2730 can be easily distinguished from neutron patterns because gamma patterns are darker, have much lower cumulative pixel intensity, do not contain any saturated pixels, and the patterns do not tend to be symmetrical. Asymmetric patterns generally tend not to be associated with neutrons. Due to the stark differences in features between gamma and neutron patterns at the pixel level, deep learning software can be easily trained to distinguish between the two types of patterns.

[0247] Figure 28 illustrates pixel-level images with neutron counts under high gamma conditions, according to an embodiment of the present invention. The pixel-level images shown in Figure 28 each include bright neutron counts 2890 under 1000 Gy / hour gamma conditions. As shown in Figure 28, at 1000 Gy / hour, the bright counts 2890 in each frame are less visible than at 62 Gy / hour, but are still reliably distinguishable by shape (size and symmetry) and intensity. Thus, deep learning algorithms can be easily trained to recognize bright neutron counts even in images with high levels of gamma radiation, e.g., 1200 Gy / hour.

[0248] FIG. 29 shows a flowchart 2900 of an exemplary computer-implemented process for detecting the presence of neutrons in an image generated from sensor information in accordance with an embodiment of the present invention.

[0249] In step 2902, first a neutron pattern needs to be established based on prior knowledge of the physics of neutron behavior and other information, such as the known symmetric behavior of neutron counts.

[0250] A collection of training images containing known neutron patterns within them are labeled in step 2904. In other words, the neutron patterns in the training images are identified and pixels that are part of the neutron patterns are unambiguously labeled.

[0251] In step 2906, deep learning software, for example a deep CNN, is trained to recognize known patterns using the labeled training images.

[0252] In step 2908, the artificial intelligence software is tested on a new set of images known to have neutrons present or absent. For example, a number of images known to contain neutrons or to have no neutrons are fed to the artificial intelligence software.

[0253] If the accuracy exceeds a certain threshold, new images are ready to be tested using the deep learning software, step 2910. If the accuracy is not satisfactory, the software needs to be retrained with more images, step 2912.

[0254] As previously described, in one embodiment, deep learning software can be programmed to run on computer modules 2130A-2130D shown in Figure 21. In one embodiment, once the artificial intelligence software detects a neutron, information received from the command and control module can be used to precisely identify the sensor number, frame number, timestamp, and x,y coordinates of the pixel where the neutron was detected. In one embodiment, the information, along with an image containing the neutron pattern, is displayed on display 2150 for a user to view.

[0255] In one embodiment, software running on the computational module can also be programmed to combine neutron detection information from various sensors, for example, all sensors arranged in a cube configuration as shown in FIG. 13 , to determine the source of the neutrons using the triangulation method described above. Additionally, in applications such as the damaged nuclear reactor described above, a robot sent into the containment unit to perform a scan (e.g., equipped with the exemplary detector unit from FIG. 14 ) can be programmed to operate in an autonomous mode and independently locate the source of the neutrons using information from the triangulation. In one embodiment, machine learning can also be used to teach the robot to independently locate the source of the neutrons within the containment unit. The robot can also use the information collected from the various sensors to generate a 3D debris map based on the detected neutrons.

[0256] FIG. 30 shows a flowchart 3000 of an exemplary computer-implemented process for analyzing an image to detect neutrons using a deep learning process in accordance with an embodiment of the present invention.

[0257] In step 3002, the input image is divided into multiple frames or patches. For example, given an input image of a certain dimension (such as 752 x 480 pixels) captured from a sensor, software running on computation modules 2130A-2130D first divides the image into frames or patches (e.g., 48 x 48 or 96 x 96 pixels).

[0258] In step 3004, these patches are passed through a trained deep learning model (e.g., a deep CNN) using a forward pass.

[0259] In step 3006, a probability value is received from the deep learning process. In other words, the output from the deep CNN is the probability that each image pixel is a neutron. In one embodiment, a predetermined threshold value may be used above which a pixel is considered to be a neutron pixel. For example, a pixel may be considered to be a neutron pixel above a predetermined threshold of 75% probability. Thereafter, if the deep CNN indicates that the pixel exceeds the 75% threshold probability value, the pixel is considered to be a neutron-related pixel.

[0260] In step 3008, the image is stitched back together to restore the original image dimensions.

[0261] In step 3010, the total number of neutrons present in the stitched image is counted using the connected components for each pixel. For example, as seen in Figures 26A, 26B, and 27A, neutrons are identified not only by considering saturated or high-intensity pixels, but also by examining groups of pixels surrounding high-intensity pixels. Thus, to identify neutrons, all pixels adjacent to a pixel identified as exceeding a threshold level of probability are examined. The examined connected components generally neighbor, share a vertex with, or share a common boundary with the identified pixel of interest. Additionally, any pixel identified as a connected component with a relatively higher intensity may be examined. By examining the characteristics of the pixels of interest along with the connected pixels, the deep learning process can determine whether the cumulative pattern formed by the pixel of interest and the connected pixels is associated with neutrons or gamma.

[0262] In step 3012, neutron locations are determined using a binary mask. The binary mask is a control image with the same number of pixels as the image that was passed through the deep learning process. However, in the binary mask, all pixels that do not have any part of the neutron-generated pattern may be assigned a zero, and all pixels that have any part of the neutron-generated pattern may be assigned a one. Once the binary mask is generated, it can be used to extract all neutrons from the image by convolving (pixel-by-pixel multiplication) it with the original image.

[0263] In step 3014, gamma flux in the image is measured using pixel summation. As described in connection with FIG. 27A, one way gamma radiation can be distinguished from neutrons is by summing pixel intensities. Gamma radiation generally exhibits a lower cumulative pixel intensity in a pattern than in a neutron pattern. As described in connection with FIG. 27A, summing pixel intensities involves summing the intensity values ​​of each pixel that is part of the pattern. Each pixel in a CMOS sensor has an intensity value that ranges from 0 to 255 (for an 8-bit CMOS sensor) or from 0 to 1024 (for a 10-bit CMOS sensor), etc. When gamma radiation impinges on a pixel, an electric charge is generated as a result of Compton scattering (the interaction of high-energy gamma photons with silicon). It should be noted that gamma dose rate can be reliably measured by measuring the degree of energy deposited on the pixel, which generally translates to a convolution or summation of the pixel intensities produced by the sensor in response to an independently calibrated and measured gamma field.

[0264] Visible light with photons, such as green photons, generally excites silicon electrons and moves them into the conduction band. Gamma photons contain much more energy than green photons, e.g., over a million times more. When gamma photons enter a silicon CMOS sensor pixel, they may pass through without interacting. However, if they interact, they will dislodge electrons from the atoms themselves. The gamma photons have excited the silicon electrons to such an extent that the electrons move beyond the conduction band and are dislodged from the atoms themselves. These high-energy, high-speed electrons are responsible for the charge observed in gamma-related patterns on CMOS sensors.

[0265] Because CMOS sensors are excellent charge-detecting devices, they register the charge generated in each pixel as an "intensity" value, which is a measure of how many electrons were collected in the pixel. When electrons fill a pixel, they can spill over into neighboring pixels through diffusion. The more gamma photons that strike the silicon, the more charge is generated, and the greater the intensity observed in the pixel that sees such an interaction. Therefore, a linear calibration regime for gamma incident dose can be derived, where dose is the amount of energy deposited in the silicon. A higher intensity value for a pixel in a gamma pattern simply means a higher dose of gamma radiation. From the intensity values ​​of pixels in a gamma pattern, the gamma flux can be determined. As mentioned above, embodiments of the present invention can detect neutrons in environments with gamma flux as high as 1200 Gy / hr.

[0266] In step 3016, the binary mask can also be multiplied with the input image to determine the neutron intensity. The multiplication is simply a pixel-by-pixel multiplication of two images. In other words, it is an element-by-element multiplication of two matrices, one of which has intensity values ​​in it (the original image) and the other of which has ones and zeros (the binary mask).

[0267] Finally, in step 3018, statistics for each input frame are generated and written to a file to be used for later analysis.

[0268] 31 shows a flowchart 3100 of an exemplary computer-implemented process for triangulating the source location of neutron particles, in accordance with an embodiment of the present invention. It should be noted that although FIG. 31 is described in the context of neutrons, similar techniques may be applied to triangulate the source of any particle of interest to a detector user.

[0269] In step 3101, sensing elements within the detector are geometrically configured in a variety of different ways to scan an area to find sources of radiation. For example, sensing elements or sensors can be arranged around a cube, rectangular prism, sphere, icosahedron, etc. Each of these configurations is reminiscent of multiple "eyes" scanning in some or all directions, searching for neutrons and other subatomic particles.

[0270] In step 3102, the area under investigation is inspected using a detector. In other words, the detector is used to find particles of interest within an area suspected of containing a source of radiation.

[0271] In step 3103, information is extracted from the sensors in the detector and analyzed using software tools in the command, control, and computation module shown in Figure 21. For example, the pattern of neutron impingement on the detector may be analyzed by the software. For example, if the detector is a cubed detector, as described above in connection with Figures 12 and 13, the pattern of neutron impingement on each sensor on each face of the cube may be analyzed.

[0272] As mentioned earlier, neutrons incident on one face of the cube are mostly detected by the sensing elements on that face. Because the sensing elements can be pixelated, with each pixel acting as a detection element, the angle between the sensing element and the source also creates a gradient of detection within the pixel. Within the same sensing element (e.g., the same SPU), pixels closest to the source are more likely to detect more neutrons than pixels farther away. By using the counts and profiles of neutrons detected on each side of the cube, it becomes possible to determine the location of the source by triangulating the results from all the elements on different sides of the cube.

[0273] Thus, in step 3104, by triangulating the results from all sensors on multiple surfaces of the detector, the source of radiation can be determined.

[0274] FIG. 32 shows a flowchart 3200 of an exemplary computer-implemented process for independently controlling sensors to ensure reliability, according to an embodiment of the present invention.

[0275] In step 3201, data is received from multiple sensors in a detector in one or more command and control modules 2120A-2120D.

[0276] In step 3202, based on the received data, the command and control module can determine the number of particles of interest in the detected environment. For example, if a large number of neutrons are detected in the reactor, the command and control module can determine that the signal level is high. Alternatively, the command and control module can use particle counting to determine if the signal level is low.

[0277] In step 3203, in response to determining that the signal level is high, one or more sensors in the detector are turned off. Since multiple redundant sensors are used in the detector to ensure reliability, this does not affect the overall performance of the detector. Similarly, in step 3204, in response to determining that the signal level is low, some of the redundant sensors that were previously turned off can be turned back on again.

[0278] FIG. 33 shows a flowchart 3300 of an exemplary computer-implemented process for gathering information from a tunable sensor used for particle detection in accordance with an embodiment of the present invention.

[0279] In step 3301, information is extracted from one or more tunable sensors used for particle detection. In one embodiment, the sensors can be in hostile environments. In step 3302, data from the sensors is serialized so that it can be transmitted over long distances using thin cables, e.g., cables with a diameter of 3 mm or less. In one embodiment, data from the sensors can also be tagged with the sensor's ID number and a timestamp before transmission.

[0280] In step 3303, data is transmitted over cables at frequencies in the MHz to GHz range. In one embodiment, the thin cables can be twisted pair cables, coaxial cables, or fiber optic cables. In another embodiment, the thin cables include protective shielding to ensure they can withstand adverse environmental conditions.

[0281] In step 3304, the serialized data can be received at the command and control modules 2120A-2120D, and the serialized data can be converted into a sequence of images at a rate of 60 frames per second.

[0282] Finally, in step 3305, computational modules 2130A to 2130D can be used to perform a particle discrimination procedure on the sequence of images, for example using a deep learning process, to detect the presence of particles of interest, e.g., gamma photons, neutrons, etc.

[0283] FIG. 34 shows a flowchart 3400 of an exemplary computer-implemented process for disabling non-functioning sensors to ensure detector reliability and extend the detector's operational life, in accordance with an embodiment of the present invention.

[0284] A pattern is transmitted from the command and control modules 2120A-2120D to multiple sensors in step 3401. For example, in one embodiment, a checkerboard pattern can be transmitted to multiple sensors.

[0285] In step 3402, the patterns returned from each of a plurality of sensors are received at a command and control module.

[0286] In step 3403, for each of the sensor modules, the received pattern is examined and compared to the transmitted pattern to determine deviations between the two.

[0287] In response to determining that the deviation exceeds a predetermined threshold for a particular sensor, the malfunctioning sensor is taken offline in step 3404. In other words, the command and control module has individual control over the sensors and can deactivate the malfunctioning sensor.

[0288] Additionally, in step 3405, previously deactivated sensors that were turned off (for power-related or other reasons) but are still operational may be turned back on to replace sensors that have been taken offline.

[0289] FIG. 35 shows a flowchart 3500 of an exemplary computer-implemented process for conserving power and managing heat in a tunable detector system in accordance with an embodiment of the present invention.

[0290] In step 3501, temperature readings are received in command and control modules 2120A-2120D from multiple sensors within the detector.

[0291] In step 3502, the command and control module determines whether the temperature of any of the sensors exceeds a predetermined threshold.

[0292] Sensors whose temperatures exceed a threshold are placed in partial shutdown mode in step 3503. If partial shutdown mode is not available, the sensors are shut off completely.

[0293] In step 3504, the shut down sensors are cyclically turned back on after a predetermined amount of time has elapsed. In this manner, power is conserved and heat is managed by cyclically turning sensors on and off as needed based on temperature readings from the sensors.

[0294] Although the above disclosure describes various embodiments using specific block diagrams, flowcharts, and examples, each block diagram component, flowchart step, operation, and / or component described and / or illustrated herein may be implemented individually and / or collectively using a wide variety of hardware, software, or firmware (or any combination thereof) configurations. Additionally, any disclosure of components contained within other components should be considered an example, as many other architectures can be implemented to achieve the same functionality.

[0295] The process parameters and order of steps described and / or illustrated herein are given by way of example only. For example, although the steps illustrated and / or described herein may be shown or described in a particular order, these steps do not necessarily have to be performed in the order illustrated or described. The various exemplary methods described and / or illustrated herein may omit one or more of the steps described or illustrated herein or may include additional steps in addition to those disclosed.

[0296] Although various embodiments are described and / or illustrated herein in the context of a fully functional computing system, one or more of these exemplary embodiments may be distributed as a program product in various forms, regardless of the particular type of computer-readable medium actually used to effect the distribution. The embodiments disclosed herein may be implemented using software modules that perform certain tasks. These software modules may include scripts, batch files, or other executable files that may be stored on a computer-readable storage medium or within a computing system. These software modules may configure a computing system to execute one or more of the exemplary embodiments disclosed herein. One or more of the software modules disclosed herein may be implemented in a cloud computing environment. A cloud computing environment may provide various services and applications over the Internet. These cloud-based services (e.g., software as a service, platform as a service, infrastructure as a service, etc.) may be accessible through a web browser or other remote interface. Various functions described herein may be provided through a remote desktop environment or any other cloud-based computing environment.

[0297] The foregoing description has been set forth with reference to specific embodiments for purposes of explanation. However, the illustrative description above is not intended to be exhaustive or to limit the invention to the precise form disclosed. Many modifications and variations are possible in light of the above teachings. The embodiments have been chosen and described to best explain the principles of the invention and its practical application, and thereby enable those skilled in the art to best utilize the invention and its various embodiments, with various modifications as may be suited to the particular use contemplated.

[0298] Embodiments in accordance with the present invention are thus described. While the present disclosure has been described in terms of particular embodiments, it should be understood that the present invention should not be construed as limited by such embodiments, but rather should be construed according to the following claims.

Claims

1. 1. A computer-implemented method for detecting neutrons in an image from an adjustable sensor system, comprising: training a deep learning process to recognize known radiation-dependent signature patterns produced by neutrons in test images; Dividing an input image into a number of frames; passing the plurality of frames through the deep learning process to recognize neutrons in the plurality of frames; recombining the frames to reconstruct the input image; For each pixel in the input image, examining pixels connected to the respective pixel to determine whether a signature pattern characteristic of neutrons is present in the input image; using results from the inspection to count the number of neutrons in the input image; A method comprising:

2. determining a binary mask for the input image, the binary mask assigning a value of "1" to each pixel in the image associated with the signature pattern characteristic of neutrons and assigning a value of "0" to the remaining pixels in the image; convolving the binary mask with the input image to determine neutron intensity values ​​for neutrons in the input image; The method of claim 1 further comprising:

3. 3. The method of claim 2, further comprising measuring gamma flux in the input image prior to the convolution using a summation of pixel intensities in the input image, the gamma flux having a lower cumulative intensity value than the signature pattern characteristic of neutrons.

4. generating statistics about the input image, including information about the number of neutrons, the neutron intensity values, and the gamma flux; saving said statistics for further analysis; 4. The method of claim 3, further comprising:

5. The step of training the deep learning process includes: labeling neutrons in the first plurality of images based on the known radiation-dependent signature patterns produced by the neutrons; training the deep learning process to recognize the known radiation-dependent signature pattern produced by neutrons using the labeled neutrons in the first plurality of images; testing the deep learning process to recognize the presence of neutrons using a second plurality of images, the second plurality of images being test images with known results; responsive to determining that the accuracy of recognition of the second plurality of images is above a threshold, inputting a new image to be tested into the deep learning process; training the deep learning process with additional images in response to determining that the accuracy of recognition is below the threshold; 2. The method of claim 1, comprising:

6. 10. The method of claim 1, wherein the deep learning process comprises a deep fully convolutional neural network (CNN) process.

7. 2. The method of claim 1, wherein the pixels connected to each pixel include pixels that share a common boundary or vertex with the respective pixel.

8. The step of examining pixels connected to each pixel comprises: determining the number of saturated pixels in the pattern formed by each pixel and the pixels connected to each pixel; determining whether the pattern is symmetric; determining whether the pixels in the pattern exhibit pixel intensities that decrease gradually from the center of the pattern; 2. The method of claim 1, comprising:

9. 1. A non-transitory computer-readable storage medium having stored thereon computer-executable instructions that, when executed by a computer system, cause the computer system to perform a method for detecting particles of interest in an image from a tunable sensor system, the method comprising: training a deep learning process to recognize known radiation-dependent signature patterns produced by particles of interest in test images; Dividing an input image into a number of frames; passing the plurality of frames through the deep learning process to recognize the particle of interest in the plurality of frames; combining the frames to reconstruct the input image; For each pixel in the input image, examining pixels connected to the respective pixel to determine whether a signature pattern characteristic of the particle of interest is present in the input image; determining a count of the particles of interest in the input image using the connected pixels; 1. A non-transitory computer-readable storage medium comprising:

10. The method comprises: determining a binary mask for the input image, the binary mask assigning a value of "1" to each pixel in the image associated with the signature pattern unique to the particle of interest and assigning a value of "0" to the remaining pixels in the image; convolving the binary mask with the input image to determine an intensity value for the particle of interest in the input image; 10. The non-transitory computer-readable storage medium of claim 9, further comprising:

11. The method comprises: generating statistics about the input image including information about the count of the particles of interest and the intensity values ​​of pixels associated with the signature pattern; saving said statistics for further analysis; 11. The non-transitory computer-readable storage medium of claim 10, further comprising:

12. The step of training the deep learning process includes: labeling the particle of interest in a first plurality of images based on the known radiation-dependent signature pattern produced by the particle of interest; training the deep learning process to recognize the known radiation-dependent signature pattern produced by the particle of interest using the labeled particles of interest in the first plurality of images; testing the deep learning process to recognize the presence of the particle of interest using a second plurality of images, the second plurality of images being test images with known results; responsive to determining that the accuracy of recognition of the second plurality of images is above a threshold, supplying a new image to be tested to the deep learning process; training the deep learning process with additional images in response to determining that the accuracy of recognition is below the threshold; 10. The non-transitory computer-readable storage medium of claim 9, further comprising:

13. 13. The non-transitory computer-readable storage medium of claim 12, wherein the deep learning process comprises a deep fully convolutional neural network (CNN) process.

14. 10. The non-transitory computer-readable storage medium of claim 9, wherein the particles of interest are neutrons.

15. 1. A system for detecting neutrons in an image from an adjustable sensor system, comprising: a memory for storing a plurality of test images, an input image, and instructions associated with a deep learning process and a process for detecting particles of interest in the images; a processor, coupled to the memory, training a deep learning process to recognize known radiation-dependent signature patterns produced by particles of interest in the test images; Dividing an input image into a plurality of frames; passing the plurality of frames through the deep learning process to recognize the particle of interest in the plurality of frames; combining the frames to reconstruct the input image; For each pixel in the input image, examining pixels connected to the respective pixel to determine whether a signature pattern characteristic of the particle of interest is present in the input image; determining a count of the particles of interest in the input image using the connected pixels; the processor configured to operate in accordance with the instructions to perform A system comprising:

16. The processor: determining a binary mask for the input image, the binary mask assigning a value of "1" to each pixel in the image associated with the signature pattern characteristic of the particle of interest and assigning a value of "0" to the remaining pixels in the image; convolving the binary mask with the input image to determine an intensity value for the particle of interest in the input image; 16. The system of claim 15, further configured to:

17. The processor: generating statistics about the input image including information about the count of the particles of interest and the intensity values ​​of pixels associated with the signature pattern; Saving the statistics for further analysis.

17. The system of claim 16, further configured to:

18. To train the deep learning process, the processor: labeling the particle of interest in a first plurality of images based on the known radiation-dependent signature pattern produced by the particle of interest; training the deep learning process to recognize the known radiation-dependent signature pattern produced by the particle of interest using the labeled particles of interest in the first plurality of images; testing the deep learning process to recognize the presence of the particle of interest using a second plurality of images, the second plurality of images being test images with known results; responsive to determining that the accuracy of recognition of the second plurality of images is above a threshold, supplying a new image to be tested to the deep learning process; training the deep learning process with additional images in response to determining that the accuracy of recognition is below the threshold; and 16. The system of claim 15, configured to:

19. 20. The system of claim 18, wherein the deep learning process comprises a deep fully convolutional neural network (CNN) process.

20. The system of claim 15, wherein the particles of interest are neutrons.

Images