Method and system for performing instant automatic calibration adjustment of

Abstract

The present specification discloses a method of adjusting calibration data in an X-ray inspection system. Calibration data is initially generated. An X-ray scan image of the cargo container is then acquired. Each X-ray scan image is segmented into regions of interest, wherein the regions of interest enclose a known material or a material corresponding to a known HS code in volume. First data indicative of Zeff for each of the regions of interest is determined using the calibration data. The first data is compared to second data indicative of a known Zeff corresponding to a known material and / or HS code. Then, if the first data and the second data are significantly different, the calibration data is adjusted to generate second calibration data. The calibration data is replaced with the second calibration data in the memory.

Description

[0001] Cross-references

[0002] This application relies on the priority of U.S. Patent Provisional Application No. 63 / 369,386, filed on July 26, 2022, entitled “Methods and Systems for Performing On-The-Fly Automatic Calibration Adjustments of X-Ray Inspection Systems,” the entire contents of which are incorporated herein by reference. Technical Field

[0003] The present specification relates generally to the field of X-ray inspection. More particularly, the present specification relates to a method of automatically adjusting calibration data of an X-ray inspection system while scanning an object under inspection or examination. Background Art

[0004] Most modern high energy (2.5-9 MeV) transmission X-ray non-intrusive inspection systems provide some degree of cargo composition estimation by using staggered dual energy pulses. By comparing low energy (3-5 MeV) and high energy (6-9 MeV) radiographic images of the same container, different areas of the container interior can be characterized into four major categories of materials: organic materials, inorganic materials (such as soil, fertilizer, salt, aluminum), metallic materials (such as steel, scrap metal), and high Z materials (such as materials with atomic numbers equal to or higher than tungsten).

[0005] Unlike low-energy inspection systems used for package and baggage screening, where the dominant process is the photoelectric effect, at high energies the Compton effect dominates the interaction of X-rays with matter. The Compton process is weakly dependent on the effective atomic number of the material, making it difficult to make an accurate assessment.

[0006] In addition to the challenges posed by physical properties, there are other challenges specific to cargo screening, such as, but not limited to: a) cargo clutter, where multiple optically dense objects overlap in the image; b) X-ray source stability (in terms of maintaining consistent and / or predictable values ​​for dose and energy) and detector stability, both of which can cause calibration drift; c) energy-dependent X-ray scatter, where low-energy and high-energy beams produce different blurs in the image, and d) obscured objects of interest (such as high-Z shields, weapons, or smuggled pharmaceutical packages), which are typically small compared to the rest of the imaged cargo.

[0007] The dual energy material separation method involves a calibration process where materials with known atomic number and density (such as plastic, aluminum, steel, lead, tungsten) of varying thickness are placed between the source and the detector, a scan is performed and the detector response to the high and low energy beam components is recorded. In normal operating mode, the reverse process occurs. The detector records the high and low energy responses to the object in the beam, and then the algorithm determines the closest candidate in terms of material type and thickness based on the effective atomic number (Zeff). Uncertainties are associated with the determination of the material type, which depends on the system geometry, the X-ray flux and energy, and the number of samples (pulse pairs) used to make the determination.

[0008] Unfortunately, material separation calibration does not remain stable over the life of an X-ray inspection system. Changes often occur that make the Zeff calculation unreliable. Under some conditions, drift may even occur between two scans performed a few minutes apart.

[0009] Therefore, there is a need for a method of automatically modulating the calibration data of an X-ray system during scanning of an object under inspection or examination. Summary of the invention

[0010] The following embodiments and aspects thereof are described and illustrated in conjunction with systems, tools, and methods, which are intended to be exemplary and illustrative, not limiting in scope.This application discloses many embodiments.

[0011] In some embodiments, the present specification discloses a method for performing on-the-fly adjustment of calibration data corresponding to an X-ray inspection system, the X-ray inspection system including an X-ray source in data communication with a computing device, the method being implemented in a computing device having one or more physical processors programmed with a plurality of program instructions, the plurality of program instructions causing the computing device to perform the method when executed by the one or more physical processors, the method comprising: generating a plurality of first calibration data by the X-ray inspection system, wherein the plurality of first calibration data is stored in a memory associated with the computing device; generating data representing an X-ray scan image of a cargo container using the X-ray inspection system; and segmenting the data representing the X-ray scan image. The invention relates to a method for forming a plurality of regions of interest, wherein at least some of the plurality of regions of interest contain only one type of material having at least one of a known effective atomic number (Zeff) or a known HS code; determining first data using a first plurality of calibration data, the first data indicating the Zeff of at least some of the plurality of regions of interest; comparing the first data with second data, the second data indicating the Zeff corresponding to the known Zeff or the known HS code; adjusting the first plurality of calibration data to generate a second plurality of calibration data if the first data and the second data differ by more than a predefined threshold; and processing data representing an X-ray scan image using the second plurality of calibration data to generate and display the X-ray scan image.

[0012] Optionally, the method also includes calculating the Zeff value of each pixel in the X-ray scan image; applying a smoothing algorithm to the Zeff values ​​of multiple pixels associated with each of a plurality of regions of interest to generate a smoothed Zeff-based region of interest; applying a predefined color map to the smoothed Zeff-based region of interest; and displaying the color-coded X-ray scan image.

[0013] Optionally, if the first data and the second data differ by more than a predefined threshold, the second plurality of calibration data are used to calculate the Zeff value of each pixel, and if the first data and the second data differ by no more than a predefined threshold, the first plurality of calibration data are used to calculate the Zeff value of each pixel.

[0014] Optionally, the smoothing algorithm comprises any one of a simple sliding average, a shape preserving smoothing filter, image guided segmentation or K-means clustering.

[0015] Optionally, at least some of the plurality of regions of interest correspond to empty or unobstructed areas of the cargo container.

[0016] Optionally, at least some of the plurality of regions of interest correspond to at least one of a volume surrounding only a top surface of the cargo container, a volume surrounding only a front end of the cargo container, or a volume surrounding only a rear end of the cargo container.

[0017] Optionally, the first plurality of calibration data is generated by placing a set of materials with known atomic numbers and densities between a radiation source and a detector of the X-ray inspection system, wherein the set of materials has different thicknesses; and recording the response of the detector to each type of material in the set of materials.

[0018] Optionally, the X-ray scan images correspond to low energy and high energy of the radiation source.

[0019] In some embodiments, the present specification relates to an X-ray inspection system, comprising: a radiation source; a detector array; a computing device, the computing device being in data communication with the radiation source and the detector array, wherein the computing device has one or more physical processors programmed with a plurality of program instructions, the plurality of program instructions, when executed by the one or more physical processors, causing the computing device to: generate a first plurality of calibration data using the X-ray inspection system, wherein the first plurality of calibration data is stored in a memory associated with the computing device; generate an X-ray scan image of a cargo container using the X-ray inspection system; segment the X-ray scan image into a plurality of regions of interest, wherein each of the plurality of regions of interest contains a single type of material associated with a known atomic number or a single type of material associated with a known atomic number; HS code; determining first data using a first plurality of calibration data, the first data indicating a first effective atomic number of the single type of material associated with the known atomic number, or a second effective atomic number of the single type of material associated with the known HS code, for each of the plurality of regions of interest; comparing the first data with second data, the second data indicating a known effective atomic number corresponding to the material associated with the known atomic number, or a known effective atomic number corresponding to the material associated with the known HS code; if the first data and the second data are significantly different, adjusting the first plurality of calibration data to generate a second plurality of calibration data; and replacing the first plurality of calibration data with the second plurality of calibration data in a memory.

[0020] Optionally, the plurality of program instructions, when executed by one or more physical processors, further cause the computing device to: calculate an effective atomic number value for each pixel in an X-ray scan image; apply a smoothing algorithm to the effective atomic number values ​​of a plurality of pixels associated with each of a plurality of regions of interest so as to generate smoothed effective atomic number-based regions of interest; apply a predefined color map to the smoothed effective atomic number-based regions of interest; and display a color-coded X-ray scan image.

[0021] Optionally, if the first data and the second data are significantly different, the second plurality of calibration data are used to calculate the effective atomic number value for each pixel, and if the first data and the second data are not significantly different, the first plurality of calibration data are used to calculate the effective atomic number value for each pixel.

[0022] Optionally, the smoothing algorithm comprises any one of a simple running average, a shape preserving smoothing filter, image guided segmentation or K-means clustering.

[0023] Optionally, at least some of the plurality of regions of interest are empty or unobstructed.

[0024] Optionally, at least some of the plurality of regions of interest contain at least one of a volume surrounding only a top surface of the cargo container, a volume surrounding only a front end of the cargo container, or a volume surrounding only a rear end of the cargo container.

[0025] Optionally, the first plurality of calibration data is generated by placing a set of materials with known atomic numbers and densities between a radiation source and a detector of the X-ray inspection system, wherein the set of materials has different thicknesses; and recording the response of the detector to each type of material in the set of materials.

[0026] Optionally, the X-ray scan images correspond to low energy and high energy of the radiation source.

[0027] The foregoing and other embodiments of the present specification are described in more depth in the drawings and detailed description provided below. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The accompanying drawings illustrate various embodiments of systems, methods, and embodiments of various other aspects of the present disclosure. It will be understood by those of ordinary skill in the art that the element boundaries (e.g., boxes, box groups, or other shapes) shown in the figures represent an example of boundaries. It is possible that, in some examples, an element may be designed as multiple elements, or multiple elements may be designed as one element. In some examples, an element shown as an internal component of an element may be implemented as an external component in another element, and vice versa. In addition, the elements may not be drawn to scale. A non-restrictive and non-exhaustive description is described with reference to the following drawings. The components in the drawings are not necessarily drawn to scale, but the emphasis is placed on the explanation principle.

[0029] FIG. 1 is a schematic diagram of an exemplary LINAC-based high-energy X-ray cargo inspection system that can implement the methods described in this specification in an embodiment;

[0030] FIG2A illustrates segmented regions of a radiographic scan image of a cargo container according to some embodiments of the present specification;

[0031] FIG. 2B shows a radiographic scan image of a container containing bananas, according to some embodiments of the present specification; and

[0032] 3 is a flow chart of exemplary steps of a method of processing an X-ray scan image of an object under inspection for material classification and on-the-fly automatic adaptive calibration, according to some embodiments of the present specification. DETAILED DESCRIPTION

[0033] This specification relates to multiple embodiments. The following disclosure is provided to enable one of ordinary skill in the art to practice the present invention. The language used in this specification should not be interpreted as a general negation of any one specific embodiment or used to limit the meaning of the claims beyond the terms used therein. Without departing from the spirit and scope of the present invention, the general principles defined herein can be applied to other embodiments and applications. In addition, the terms and wordings used are for the purpose of describing exemplary embodiments and should not be considered restrictive. Therefore, the present invention will be given the widest scope including many substitutions, modifications and equivalents consistent with the disclosed principles and features. For the sake of clarity, details related to technical materials known in the technical field related to the present invention are not described in detail to avoid unnecessarily obscuring the present invention.

[0034] In various embodiments, the computing device includes an input / output controller, at least one communication interface, and a system memory. The system memory includes at least one random access memory (RAM) and at least one read-only memory (ROM). These elements communicate with a central processing unit (CPU) to implement the operation of the computing device. In various embodiments, the computing device can be a conventional stand-alone computer, or alternatively, the functionality of the computing device can be distributed across multiple computer systems and architectures.

[0035] In some embodiments, the execution of multiple program instructions or code sequences enables the CPU of the computing device to perform various functions and processes. In alternative embodiments, hard-wired circuits can be used instead of software instructions or in combination with software instructions to implement the processes of the systems and methods described in this application. Therefore, the described systems and methods are not limited to any specific combination of hardware and software.

[0036] In the specification and claims of the present application, each of the words "includes," "comprising," "having," "containing," and forms thereof is not necessarily limited to the members of the list associated with the word. Therefore, they are intended to be equivalent in meaning and open-ended, in that one or more items following any of these words is not meant to be an exhaustive list of such one or more items, or to be limited to the listed one or more items. It should be noted herein that any feature or component described in association with a particular embodiment may be used and implemented with any other embodiment unless expressly indicated otherwise.

[0037] It must also be noted that, as used herein and in the appended claims, the singular forms "a," "an," and "the" include plural referents unless the context dictates otherwise. Although any systems and methods similar or equivalent to those described herein can be used to practice or test embodiments of the present disclosure, the preferred systems and methods are now described.

[0038] “On-the-fly adjustment” means that adjustments are performed on the calibration data 1) immediately after a scan has occurred (a first scan) and before an immediately subsequent scan is performed, 2) based on scan data just acquired from the first scan, and 3) without manual or human intervention.

[0039] Overview

[0040] FIG1 shows an exemplary LINAC-based high-energy X-ray cargo inspection system that can be configured and used in embodiments to implement the methods described in this specification. As shown, the cargo inspection system 130 includes a high-energy radiation source 105 for irradiating an object 110 under inspection with a vertically divergent fan-shaped radiation beam 115. The high-energy radiation source 105 can be, but is not limited to, a linear accelerator (LINAC) or an electron beta accelerator. In embodiments, the LINAC or any other source provides a radiation dose sufficient to image containers and cargo. In embodiments, the energy and dose output of the LINAC or any other source range from 750keV to 10MeV and 0.07Gy / min to 15Gy / min, respectively.

[0041] The selection of source type, source intensity, and energy output depends on the sensitivity of the detector, the radiographic density of the cargo located in the space between the source and the detector, radiation safety considerations, and operational requirements, such as inspection speed. One of ordinary skill in the art will appreciate that the factors that need to be considered in selecting the type of radiation source depend on the inspection requirements. In one embodiment, where the object 110 under inspection is a large-sized container or automobile that highly attenuates the X-ray beam, the radiation can come from an X-ray source operating at energies ranging from about 750 keV to as high as 10 MeV or more. In various embodiments, the object 110 under inspection can be a vehicle, truck, rail car, or other container used to carry cargo, passenger luggage, or general items.

[0042] The cargo inspection system 130 also includes a detector array 120, which is preferably positioned behind the object under inspection 110 and is configured to detect radiation transmitted through the object under inspection 110. In an embodiment, the detector 120 is formed by a crystal stack that generates an analog signal when an X-ray impinges thereon, wherein the signal strength is proportional to the amount of beam attenuation in the object under inspection 110. In one embodiment, the X-ray beam detector arrangement consists of a linear array of solid-state detectors of the crystal diode type. A typical arrangement uses a cadmium tungstate scintillator crystal to absorb X-rays transmitted through the object under inspection 110 and convert the absorbed X-rays into photons of visible light. As known to those of ordinary skill in the art, crystals such as bismuth germanate, sodium iodide or other suitable crystals may be used instead. The crystal may be directly coupled to a suitable detector, such as a photodiode or a photomultiplier. The detector photodiode may be arranged linearly, which provides advantages over photomultipliers in terms of operating range, linearity and detector-to-detector matching through a unity gain device. In another embodiment, a partitioned detector is used as an alternative to a linear array detector. Such a partitioned detector may be a scintillating strip, such as cesium iodide or other material known in the art, viewed by a suitable camera or a camera optically coupled to a charge-coupled device (CCD).

[0043] It will be apparent to those skilled in the art that the cargo inspection system 130 shown in FIG1 is only one example of an inspection system that employs a high energy X-ray source such as, but not limited to, a LINAC or a betatron. In some embodiments, the radiation source 105 uses interleaved dual energy pulses to generate low energy (3-5 MeV) and high energy (6-9 MeV) X-ray scan images of the object 110 under inspection.

[0044] Calibration drift

[0045] This specification recognizes that calibration data for material separation does not remain stable during the life of the X-ray inspection system. Changes in calibration (also known as "calibration drift") often occur and make the Zeff (effective atomic number) calculation unreliable for effective material separation data. Under some conditions, drift may even occur between two scans performed a few minutes apart. The following sources, factors, causes or variables may all contribute to calibration drift. The X-ray source itself may be a contributing cause of calibration drift. Dose drift can be caused by changes in beam current or beam energy. These include magnet changes (betatron), automatic frequency control (AFC) shifts, and waveguide temperature changes. Beamforming mechanisms that may include collimator alignment cast shadows, detector stack boom oscillations, and detector collimator fin shifts may also contribute to calibration drift. In addition, X-ray detectors may have problems with photodiode dark current and scintillator light output, which contribute to calibration drift.

[0046] More than one cause of calibration drift may occur simultaneously. However, only a few of the sources, factors, causes or variables of calibration drift have built-in mechanisms for compensation, such as, for example, reference detectors and dark current measurements at the start of a scan or between pulses. These sources, factors, causes or variables of calibration drift are often non-linear and often linked to each other in complex ways. Furthermore, a complete recalibration is time consuming and is only a temporary solution for causes such as temperature drift or slight magnetic drift (in betatrons) or AFC drift (in linear accelerators).

[0047] Automatic adaptive calibration

[0048] In accordance with aspects of the present specification, instead of parameterizing the contribution of each source, factor, cause, or variable to calibration drift, a phenomenological model is developed wherein the system is configured to check the accuracy of the Zeff calculation in nearly every scan, and wherein the system is configured to automatically adjust the calibration data in order to improve the scan data, particularly material separation accuracy.

[0049] FIG. 2A shows segmented regions of a radiographic scanned image 200 of a cargo container 202 according to some embodiments of the present specification. In a non-limiting exemplary scenario, the image 200 has at least five regions or segments (also referred to as regions of interest or segments), namely, a first region 205, a second region 207, a third region 209, a fourth region 211, and a fifth region 213, the material of which is generally known as steel (e.g., for ISO containers). In addition, for at least five regions 205, 207, 209, 211, 213, the attenuation and angular range (relative to the central axis of the X-ray beam used to scan the cargo container 202) are sufficiently varied. In an embodiment, the regions are selected for different purposes. For example, referring to FIG. 2A, and by way of example only, regions 207, 209, and 211 are selected for stability because the variation is relatively small compared to regions 205 and 213, which are selected to have large angular coverage and varying attenuation.

[0050] In an embodiment, a radiographic scan image, such as the image 200 of the cargo container 202, is segmented into one or more regions or segments of interest based on the variability of attenuation and the angular range relative to the central axis of the X-ray beam. In some embodiments, the one or more regions or segments of interest are selected such that the expected attenuation variation (of the region or segment of interest) exceeds 10% and the angular range or span is greater than 2 degrees relative to the central axis of the X-ray beam.

[0051] Containers, such as cargo container 202, typically have one or more areas that remain unobstructed, such as, for example, the front end (including first area 205) and the rear end (including fifth area 213) and the top portion (including second area 207, third area 209, and fourth area 211). It should be noted that for most cargo containers, fifth area 213 and at least second top area 207 are almost always unobstructed. The first area or front end 205 is also unobstructed in most cases. The second area 207, the third area 209, and the fourth area 211 are illustrative, non-limiting examples of partitions or portions of the image 200 that are found to be unobstructed and can be used for calibration adjustments. However, depending on how the container is loaded, these areas will vary from case to case. In some embodiments, the bottom surface or portion 227 of the cargo container 202 (with known container materials) may also be at least partially unobstructed. Thus, in various embodiments, an X-ray scan image of a container is analyzed and segmented to identify regions or partitions of known material composition, wherein such regions or partitions of known material composition include portions inside or within the container, such as, but not limited to, unobstructed areas of the container (i.e., areas not loaded with cargo), and portions outside the container, such as, but not limited to, tires, windshields, frames around axles, and fuel or gas tanks.

[0052] For example, in one embodiment, the region of interest including the top portion of the cargo container 202 (including at least one of the second area 207, the third area 209, and the fourth area 211) is made only of a material that separates the interior volume of the cargo container 202 from the external environment. In another embodiment, the material that separates the interior volume of the cargo container 202 from the external environment constitutes at least 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or any numerical increment between 20% and 90% of the region of interest including the top portion of the cargo container 202.

[0053] In another example, in an embodiment, the region of interest including the front end of the cargo container 202 (including the first area 205) is made only of material that separates the interior volume of the cargo container 202 from the external environment. In another embodiment, the material that separates the interior volume of the cargo container 202 from the external environment constitutes at least 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or any numerical increment between 20% and 90% of the region of interest including the front end of the cargo container 202.

[0054] In yet another example, in one embodiment, the region of interest including the rear end of the cargo container 202 (including the fifth region 213) is made only of a material that separates the interior volume of the cargo container 202 from the exterior environment. In another embodiment, the material that separates the interior volume of the cargo container 202 from the exterior environment constitutes at least 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or any numerical increment between 20% and 90% of the region of interest including the rear end of the cargo container 202.

[0055] In yet another example, in one embodiment, the region of interest including the bottom surface or portion 227 of the cargo container 202 is made only of material that separates the interior volume of the cargo container 202 from the external environment. In another embodiment, the material that separates the interior volume of the cargo container 202 from the external environment constitutes at least 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or any numerical increment between 20% and 90% of the region of interest including the bottom surface or portion of the cargo container 202.

[0056] In addition, the first region 205 and the fifth region 213 are of interest because the angular coverage or span exceeds 15 degrees from the central axis of the X-ray beam. As known to those of ordinary skill in the art, due to the physics of X-ray generation, if the energy of the electrons that generate the X-rays changes, the shape of the attenuation curve is expected to change. High energy beams have a flux and energy component roll-off as a function of the angle relative to the central axis of the X-ray beam. The roll-off function is different at 4 MeV and 6 MeV. In some embodiments, changes, drifts, or modifications to the calibration curve are determined by calculating changes in the roll-off curve.

[0057] According to some embodiments, unobstructed regions such as second region 207, third region 209, and fourth region 211 are identified and used as data points. In some embodiments, machine learning techniques such as, but not limited to, Mask R-CNN (Region-based Convolutional Neural Network) or U-Net are configured to segment image 200 to obtain more features of the interior and exterior of cargo container 202 suitable for automatic calibration adjustment.

[0058] In some embodiments, for scans with HS codes associated with the X-ray image, the cargo inside the container is used to verify and adjust (if necessary) the calibration data. For example, FIG. 2B shows a radiographic scan image 220 of a container 230 containing bananas according to some embodiments of the present specification. The dashed box 225 shows that the cargo segmentation algorithm successfully identified the cargo area 222 inside the refrigerated ISO container 230.

[0059] According to various embodiments, there are a number of other sections or areas inside the container 230 that can be used to check the calibration data and, if necessary, adjust the calibration data. The top edge region 235 of the container 230 is metallic and clean, and the rear edge 240 of the container 230 is also metallic and spans a large range of angles relative to the central axis of the X-ray beam used to scan the container 230. Inside, the banana 222 is organic, with large angle and attenuation coverage.

[0060] Thus, in embodiments, the present specification relates to a radiographic or X-ray image processing method for identifying or segmenting one or more regions or sections of interest of a cargo container for which the material is known and has associated large angles and attenuation coverage. In various embodiments, the one or more regions or sections of interest have an associated angular span greater than 2 degrees and an attenuation variation greater than 10% relative to the average within each of the one or more regions or sections of interest. Thus, in embodiments, this refers to a specific value relative to the value within the region itself. As a non-limiting example, this may represent regions 205 and 213. Thereafter, the method computationally links the variation in Zeff values ​​of the identified or segmented regions of interest relative to the current calibration and adjusts the parameters of the calibration curve to match the known material.

[0061] In an embodiment, a region of interest may become obstructed by known features and materials (such as, for example, the skin of a cargo container). In order to be a region of interest, one condition is to be able to determine (or have a way to know) what material the region of interest contains. This information is obtained through the identification of known features (e.g., cargo container components, vehicle tires, windshields, etc.), HS codes, or other ways of identifying objects. For example, a machine learning algorithm can be configured to identify tires, tiles, or engine blocks inside a cargo container because these are made of known materials.

[0062] 3 is a flowchart of a plurality of exemplary steps of a method 300 for processing an X-ray scan image of an object under inspection for material classification and on-the-fly automatic adaptive adjustment of calibration data according to some embodiments of the present specification. In an embodiment, the method 300 includes a plurality of program instructions or codes stored in a non-transitory memory associated with a computing device, the computing device being in data communication with an X-ray inspection system configured to generate the X-ray scan image. The computing device includes a processor and a random access memory, wherein the processor executes the method 300.

[0063] In some embodiments, the X-ray inspection system is configured to use staggered dual energy pulses to generate low energy (3-5 MeV) and high energy (6-9 MeV) X-ray scan images of the cargo container. In various embodiments, the object under inspection may be a vehicle, truck, rail car, or other container used to carry cargo, passenger luggage, or general items. In a non-limiting exemplary scenario, the steps of method 300 are described with reference to a cargo container scanned using a dual energy X-ray inspection system. Alternative embodiments may use a single energy X-ray inspection system. In addition, embodiments may use a multi-view X-ray inspection system with single energy or dual energy capabilities.

[0064] At step 302, as a result of an initial calibration process, the X-ray inspection system is configured to generate a first plurality of calibration data. The first plurality of calibration data is stored in a non-transitory memory associated with a computing device. In the initial calibration process, a set of materials (such as plastic, aluminum, steel, lead, tungsten) with different thicknesses and known atomic numbers and densities are placed between an X-ray source (of the X-ray inspection system) and a detector, and the detector's responses to high energy beam components and low energy beam components are recorded corresponding to the first plurality of calibration data.

[0065] At step 304, an X-ray inspection system is operated to scan a cargo container, thereby generating data indicative of a dual energy X-ray scan image. At step 306, a computing device is configured to apply a cargo material segmentation algorithm known to a person of ordinary skill in the art to the X-ray scan image so as to identify and segment first zones of interest or regions of interest (ROI) having known material composition (in some embodiments, the first zones of interest or regions contain only one type of material having a known effective atomic number (Zeff)). For example, for an ISO container, the first zones of interest or regions include the top and ends (front and rear), which may be unobstructed and known to be steel. The unobstructed zones or regions may also include those portions of the container that are empty or without any cargo or items present. In further examples, the first zones of interest or regions may additionally include tires, windshields, frames around axles, and fuel tanks.

[0066] It will be appreciated that the calibration curve is non-linear.While a single region of interest or segment can be used for material classification and instant automatic adaptive adjustment of the calibration data, the greater the number of regions of interest or segments processed, the more accurate the recalibration or adjustment of the calibration data becomes.

[0067] At step 308, the computing device is configured to determine whether one or more HS (Harmonized Commodity Description and Coding System) codes associated with the cargo or articles of the cargo container, and therefore the X-ray scan image, are known. HS codes are identification codes assigned to articles for use in international trade. HS codes are managed by the World Customs Organization (WCO) and are accepted internationally by customs authorities and companies for identifying articles / cargo.

[0068] If one or more HS codes are known, then at step 310, the computing device is configured to further segment the cargo or article region of the X-ray scan image into one or more second regions or areas of interest corresponding to the one or more HS codes, and then the flow moves to step 312. At step 312, the computing device is configured to determine first data using the first plurality of calibration data, the first data indicating the Zeff of the segmented first region of interest and / or second region of interest.

[0069] If one or more HS codes are unknown, the computing device is configured to determine first data using the first plurality of calibration data, the first data indicating the Zeff of the segmented first region of interest. The second region of interest including the known HS code is used to provide additional data points to the recalibration model. As previously described, the segmented first region of interest is a region corresponding to a known material, and the segmented second region of interest is a region corresponding to one or more HS codes. Therefore, if one or more HS codes are known, the computing device is configured to determine first data using the first plurality of calibration data, the first data indicating the Zeff of the segmented first region of interest and the segmented second region of interest.

[0070] At step 314, the computing device is configured to compare the first data with the second data, the second data indicating a reference Zeff of a known material and / or HS code corresponding to the segmented first region of interest and / or second region of interest. At step 316, the computing device is configured to determine whether the first data is significantly different or deviates from the second data. In some embodiments, if the first data and the second data differ or deviate by at least 2 atomic number units, the first data is considered to be significantly different or deviated from the second data. In some embodiments, if the first data and the second data differ or deviate by at least 3 atomic number units, the first data is considered to be significantly different or deviated from the second data.

[0071] If a significant difference or deviation is determined, as described above, then at step 318, the first plurality of calibration data is adjusted or modulated based on the second data to generate a second plurality of calibration data, and at step 320, the first plurality of calibration data is replaced with the second plurality of calibration data for storage in a memory, and the process then moves to step 322.

[0072] It should be understood that the adjustment or modulation of the first plurality of calibration data is non-linear. The adjustment or modulation depends on the number of regions or segments of interest identified or available. In some embodiments, if the identified regions or segments of interest correspond only to container parts, a shift of the calibration curves (of the first plurality of calibration data) is performed to recalibrate. However, if regions or segments of interest with HS codes are also available, each calibration curve (in the first plurality of calibration data) corresponding to a specific type of material is modified or modulated.

[0073] At step 322, the computing device is configured to calculate the Zeff value for each individual pixel in the entire X-ray scan image. In an embodiment, if the first data and the second data are significantly different, the Zeff value for each pixel is calculated using the second plurality of calibration data, as defined above. However, if the first data and the second data are not significantly different, the Zeff value for each pixel is calculated using the first plurality of calibration data.

[0074] At step 324, the computing device is configured to apply a material segmentation algorithm to the X-ray image to generate a plurality of segments, and thereafter apply a smoothing or averaging filter or algorithm to the Zeff values ​​(of step 322) of a plurality of pixels associated with each of the plurality of segments to generate one or more smoothed Zeff-based segments, regions, or clusters in the X-ray image. In various embodiments, smoothing or averaging involves techniques such as, but not limited to, a simple sliding average of a shape-preserving smoothing filter, image-guided segmentation, and K-means clustering.

[0075] Thereafter, at step 326, the computing device is configured to apply a predefined color map or encoding to one or more smoothed Zeff-based regions of interest, segments, or clusters. For example, smoothed Zeff-based segments corresponding to organic materials are colored orange, metallic materials are colored blue, inorganic materials are colored green, and high-Z materials are colored red. Segments or regions with unreliable Zeff due to too little attenuation or too much noise remain uncolored (grayscale).

[0076] Finally, at step 328, the computing device is configured to render or display the color-coded X-ray image on a display.

[0077] The above examples are merely illustrations of many applications of the systems and methods of this specification. Although only a few embodiments of the present invention are described herein, it should be understood that the present invention may be implemented in many other specific forms without departing from the spirit or scope of the present invention. Therefore, the present examples and embodiments are considered to be illustrative rather than restrictive, and the present invention may be modified within the scope of the appended claims.

Claims

1. A method of performing on-the-fly adjustment of calibration data corresponding to an X-ray inspection system, the X-ray inspection system comprising an X-ray source in data communication with a computing device, the method being implemented in a computing device having one or more physical processors programmed with a plurality of program instructions that, when executed by the one or more physical processors, cause the computing device to perform the method, the method comprising: generating, by the X-ray inspection system, a first plurality of calibration data, wherein the first plurality of calibration data is stored in a memory associated with the computing device; generating data representing an X-ray scan image of a cargo container using the X-ray inspection system; segmenting data representing an X-ray scan image into a plurality of regions of interest, wherein at least some of the plurality of regions of interest contain only one type of material having at least one of a known effective atomic number (Zeff) or a known HS code; determining first data using the first plurality of calibration data, the first data being indicative of Zeff for the at least some of the plurality of regions of interest; comparing the first data with second data, the second data indicating a Zeff corresponding to the known Zeff or the known HS code; If the first data and the second data differ by more than a predefined threshold, adjusting the first plurality of calibration data to generate a second plurality of calibration data; and The data representing an X-ray scan image is processed using the second plurality of calibration data to generate and display an X-ray scan image.

2. The method according to claim 1, further comprising: Calculating the Zeff value of each pixel in the X-ray scan image; applying a smoothing algorithm to the Zeff values ​​of a plurality of pixels associated with each of the plurality of regions of interest to generate smoothed Zeff-based regions of interest; applying a predefined color map to the smoothed Zeff-based region of interest; and The color-coded X-ray scan image is displayed.

3. The method according to claim 2, wherein: If the first data and the second data differ by more than the predefined threshold, the second plurality of calibration data are used to calculate the Zeff value of each pixel, and wherein, if the first data and the second data differ by no more than the predefined threshold, the first plurality of calibration data are used to calculate the Zeff value of each pixel.

4. The method according to claim 2, wherein: The smoothing algorithm includes any one of simple sliding average, shape preserving smoothing filter, image guided segmentation or K-means clustering.

5. The method according to claim 1, wherein: The at least some of the plurality of regions of interest correspond to empty or unobstructed areas of the cargo container.

6. The method according to claim 1, wherein: The at least some of the plurality of regions of interest correspond to at least one of a volume surrounding only a top surface of the cargo container, a volume surrounding only a front end of the cargo container, or a volume surrounding only a rear end of the cargo container.

7. The method according to claim 1, wherein: The first plurality of calibration data is generated by the following steps: placing a collection of materials having known atomic numbers and densities between a radiation source and a detector of the X-ray inspection system, wherein the collection of materials has different thicknesses; and The response of the detector to each class of material in the set of materials is recorded.

8. The method according to claim 7, wherein: The X-ray scan images correspond to low energy and high energy of the radiation source.

9. An X-ray inspection system comprising: Radiation sources; a detector array; a computing device in data communication with the radiation source and the detector array, wherein the computing device has one or more physical processors programmed with a plurality of program instructions that, when executed by the one or more physical processors, cause the computing device to: generating a first plurality of calibration data using the X-ray inspection system, wherein the first plurality of calibration data is stored in a memory associated with the computing device; generating an X-ray scan image of the cargo container using the X-ray inspection system; Segmenting the X-ray scan image into a plurality of regions of interest, wherein each of the plurality of regions of interest contains a single type of material associated with a known atomic number or a single type of material associated with a known HS code; determining first data indicating a first effective atomic number of the single type of the material associated with the known atomic number or a second effective atomic number of the single type of the material associated with the known HS code for each of the plurality of regions of interest using the first plurality of calibration data; comparing the first data to second data indicating a known effective atomic number corresponding to the material associated with the known atomic number or a known effective atomic number corresponding to the material associated with the known HS code; If the first data and the second data are significantly different, adjusting the first plurality of calibration data to generate a second plurality of calibration data; and The first plurality of calibration data is replaced in memory with the second plurality of calibration data.

10. The X-ray inspection system according to claim 9, wherein: The plurality of program instructions, when executed by the one or more physical processors, further cause the computing device to: Calculating an effective atomic number value for each pixel in the X-ray scan image; applying a smoothing algorithm to the effective atomic number values ​​of a plurality of pixels associated with each of the plurality of regions of interest to generate smoothed effective atomic number-based regions of interest; applying a predefined color map to the smoothed effective atomic number-based region of interest; and The color-coded X-ray scan image is displayed.

11. The X-ray inspection system according to claim 10, wherein: If the first data and the second data are significantly different, the second plurality of calibration data are used to calculate the effective atomic number value of each pixel, and if the first data and the second data are not significantly different, the first plurality of calibration data are used to calculate the effective atomic number value of each pixel.

12. The X-ray inspection system of claim 10, wherein: The smoothing algorithm includes any one of simple sliding average, shape preserving smoothing filter, image guided segmentation, or K-means clustering.

13. The X-ray inspection system of claim 9, wherein: At least some of the plurality of regions of interest are empty or unobstructed.

14. The X-ray inspection system of claim 9, wherein: At least some of the plurality of regions of interest correspond to at least one of a volume surrounding only a top surface of the cargo container, a volume surrounding only a front end of the cargo container, or a volume surrounding only a rear end of the cargo container.

15. The X-ray inspection system of claim 9, wherein: The first plurality of calibration data is generated by the following steps: placing a collection of materials having known atomic numbers and densities between a radiation source and a detector of the X-ray inspection system, wherein the collection of materials has different thicknesses; and The response of the detector to each class of material in the set of materials is recorded.

16. The X-ray inspection system of claim 15, wherein: The X-ray scan images correspond to low energy and high energy of the radiation source.

Images