A multi-camera cooperative detection system based on hyperspectral and visible light imaging

Abstract

The utility model discloses a kind of multi-camera cooperative detection system based on hyperspectral and visible light imaging, it is related to food detection technical field.System includes material conveying belt, encoder, hyperspectral camera, visible light camera and logic controller;Encoder is installed on the transmission component of material conveying belt, generates the pulse signal synchronous with operation displacement;Logic controller receives pulse signal as reference trigger signal, and based on the signal generation has the first camera control signal and the second camera control signal with preset timing difference, respectively sent to visible light camera and hyperspectral camera, to compensate the physical interval between both, to control image acquisition in the same position of the detected material realizes space alignment.The utility model solves the problem that multiple camera independent control is difficult to accurately synchronize through hardware cooperative control framework, improves the precision, stability and reliability of detection system.

Description

Technical Field

[0001] This utility model relates to the field of food detection technology, and in particular to a multi-camera collaborative detection system based on hyperspectral and visible light imaging. Background Technology

[0002] In the field of modern industrial automated inspection, visible light imaging systems and hyperspectral imaging systems are two widely used non-destructive testing technologies. Visible light imaging systems, such as industrial cameras, are mainly used to acquire surface morphological features of the material being inspected, such as color, shape, and size. Hyperspectral imaging systems, on the other hand, can acquire information about the internal composition of the material, used to identify differences in chemical composition or internal quality. Both systems play important roles in their respective fields.

[0003] On some online inspection production lines, materials are carried through the inspection area by a material conveyor belt. To precisely control the production cycle or record the material position, sensors such as encoders are usually installed on the transmission components of the conveyor belt. Meanwhile, to ensure image quality, the image acquisition actions of the camera (whether a visible light camera or a hyperspectral camera) need to be coordinated with the production line speed.

[0004] However, in actual system integration, visible light imaging systems and hyperspectral imaging systems are usually designed and controlled as independent subsystems. They each have their own independent controllers and triggering mechanisms, and often lack a unified hardware control architecture capable of fine-grained coordination with the material conveyor control system. This decentralized control structure makes it difficult to achieve efficient and accurate collaborative work between various types of imaging devices and between them and the production line when integrating multiple different types of imaging equipment onto the same production line. This limits the improvement of the overall performance of the detection system and the further expansion of its functions. Utility Model Content

[0005] To address the aforementioned technical problems and deficiencies, the purpose of this utility model is to provide a multi-camera collaborative detection system based on hyperspectral and visible light imaging, which can achieve precise spatial alignment acquisition of materials by hyperspectral and visible light cameras based on material displacement, thereby solving the problem that multi-camera imaging systems are difficult to work collaboratively.

[0006] To achieve the above objectives, this utility model provides a multi-camera collaborative detection system based on hyperspectral and visible light imaging, characterized in that it includes: Material conveyor belts are used to transport materials to be tested. An encoder, installed on the transmission component of the material conveyor belt, is used to generate a pulse signal synchronized with the running displacement of the material conveyor belt; A hyperspectral camera is set along the conveying direction of the material conveyor belt to acquire hyperspectral images of the material being tested. A visible light camera is set along the conveying direction of the material conveyor belt to acquire visible light images of the material being detected. There is a preset physical distance between the visible light camera and the hyperspectral camera. The logic controller is connected to the encoder, the hyperspectral camera, the visible light camera, and the material conveyor belt, respectively. The logic controller is used to receive the pulse signal generated by the encoder as a reference trigger signal, and based on the reference trigger signal, generate a first camera control signal and a second camera control signal with a preset timing difference, and send the first camera control signal to the visible light camera and the second camera control signal to the hyperspectral camera to compensate for the preset physical distance, thereby controlling the image acquisition of the visible light camera and the hyperspectral camera to achieve spatial alignment at the same position of the material being detected.

[0007] This invention effectively solves the problem of independent and scattered imaging subsystems, making precise synchronization difficult, by constructing an integrated hardware collaborative control architecture. Its core technological advantage lies in using a unified logic controller that employs encoder pulse signals, strictly synchronized with the physical displacement of the material conveyor belt, as the reference clock for the entire system. Based on this displacement reference, the logic controller generates camera control signals with precise timing differences (a pulse count equivalent to the physical distance between the cameras) through internal hardware circuitry, thereby achieving spatial alignment triggering of visible light and hyperspectral cameras at different locations. This pure hardware control method, directly linked to displacement, effectively eliminates synchronization errors caused by production line speed fluctuations, providing a highly efficient, stable, and accurate multi-camera collaborative detection hardware platform, and improving the overall performance and reliability of the detection system.

[0008] In some implementations, the logic controller is provided with a high-speed counter module, which is used to count the pulse signals received from the encoder to achieve the preset timing difference.

[0009] By adopting the above technical solution and setting up a high-speed counter module in the logic controller, the most direct and efficient hardware guarantee for realizing encoder pulse-based delay is provided. This module can perform hardware-level counting of pulses with extremely high speed and accuracy, and its response speed and accuracy are far superior to software counting. This ensures the absolute reliability of the calculation benchmark (pulse count) for camera trigger delay, laying a solid hardware foundation for achieving precise spatial alignment at the nanosecond or microsecond level.

[0010] In some implementations, the logic controller is provided with a pulse delay timing module, which is used to perform a delay timing after receiving a pulse signal from the encoder, so as to achieve the preset timing difference.

[0011] The above technical solution provides another reliable hardware implementation path for achieving the preset timing difference. By setting a pulse delay timing module, the system can calculate a precise time delay value based on the real-time speed after triggering the first camera, and then hand it over to the hardware timing module for execution. This approach also avoids the uncertainty caused by software delays, ensuring the real-time performance and accuracy of the trigger signal generation. It is a flexible and effective hardware structure for realizing camera collaborative control.

[0012] In some embodiments, the system further includes a light source disposed above the material conveyor belt; the logic controller also includes a light source control output terminal, and the control input terminal of the light source is connected to the light source control output terminal.

[0013] By adopting the above technical solution, and adding a light source and corresponding light source control output terminal, lighting control is also incorporated into the unified hardware management of the logic controller. This allows the start and stop of the light source to be precisely synchronized with the camera's acquisition action, providing illumination only at the moment when imaging is needed. This structure not only greatly saves energy consumption and reduces production costs, but also reduces the heat and light decay generated by the light source during long-term operation, effectively extending the lifespan of the light source.

[0014] In some embodiments, the logic controller is further provided with a conveyor belt signal input terminal and an on-delay timing module. The conveyor belt signal input terminal is used to receive the stable operating status signal of the material conveyor belt. The on-delay timing module is used to delay sending an on-state signal to the visible light camera, the hyperspectral camera and the light source after receiving the stable operating status signal at the conveyor belt signal input terminal.

[0015] By adopting the above technical solution, and introducing a stable operating status signal for the conveyor belt and a turn-on delay timing module, a "stable start" hardware logic protection is added to the system. This ensures that detection only begins after the production line speed is stable and the light source intensity and color temperature have reached a stable state. From a hardware perspective, this avoids invalid data acquisition under unstable conditions during the initial system startup, thereby improving the effectiveness of the detection data and the automation and reliability of the entire system.

[0016] In some embodiments, the light source is a custom halogen light source, which includes a lamp housing, halogen lamp beads, and a voltage regulator module; the halogen lamp beads are disposed inside the lamp housing, and the voltage regulator module is used to supply power to the halogen lamp beads.

[0017] By employing the above technical solution, and by specifying the light source as a customized halogen light source containing a voltage regulator module, hardware assurance is provided for high-quality spectral data acquisition. The halogen lamp beads provide a continuous, broad spectrum covering the near-infrared region, while the voltage regulator module ensures that its luminous intensity is unaffected by power grid fluctuations. This stable hardware combination with matched spectral characteristics is a key prerequisite for ensuring the accuracy and repeatability of hyperspectral image data.

[0018] In some embodiments, multiple halogen lamps are arranged in a linear or circular array to form a uniform linear or planar lighting area above the material conveyor belt.

[0019] By adopting the above technical solution and optimizing the physical arrangement of halogen lamp beads (linear or circular array), the problem of illumination uniformity over a large field of view is solved. This array structure can form a uniform linear or planar illumination area above the material conveyor belt, effectively eliminating bright and dark areas in the image caused by uneven illumination. This ensures that all measured materials within the camera's field of view receive consistent illumination, thereby significantly improving imaging quality and detection accuracy.

[0020] In some embodiments, the lamp housing is a metal heat dissipation shell, the outer surface of the lamp housing is provided with heat dissipation fins, and the color temperature of the custom halogen light source is in the range of 2800K to 3200K.

[0021] By employing the above technical solution and limiting the hardware specifications of the lamp housing and color temperature, the long-term stability and applicability of the light source are further guaranteed. The physical structure of the metal heat dissipation housing with heat sink fins effectively solves the problem of high heat generation in halogen lamps, extending their service life. Furthermore, the color temperature range of 2800K to 3200K is the typical range where the energy distribution of halogen light sources is most suitable for spectral analysis, ensuring that the light source hardware itself can meet the stringent requirements of visible and near-infrared spectral detection.

[0022] In some embodiments, the system further includes a pre-sensor and a trigger sensor; the pre-sensor and the trigger sensor are arranged sequentially at intervals along the material conveying direction of the material conveyor belt; the logic controller is connected to the pre-sensor and the trigger sensor respectively.

[0023] By adopting the above technical solution and adding two hardware components, a preparatory sensor and a trigger sensor, the system is provided with an efficient triggering mode for handling discontinuous materials. This "preparatory-trigger" hardware logic allows the system to accurately sense the arrival of individual materials and uses the material's position as the "zero point" for initiating precise synchronous acquisition. This achieves an intelligent working mode of "measuring when there is material and stopping when there is no material," greatly improving the system's flexibility and operational efficiency when dealing with discrete material detection.

[0024] In some embodiments, the hyperspectral camera has a spectral range of 400nm-2500nm and a spectral resolution of 1nm-10nm.

[0025] By employing the above technical solution and clearly defining the core hardware parameters (spectral range and spectral resolution) of the hyperspectral camera, the detection capabilities and application boundaries of this system are fundamentally defined. The wide spectral range of 400nm-2500nm enables the system to powerfully detect the internal components of substances; while the spectral resolution of 1nm-10nm achieves an optimal balance between resolution and signal-to-noise ratio. This specific hardware selection is the foundation for the system's ability to surpass conventional visual inspection and perform in-depth quality analysis. Attached Figure Description

[0026] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments conforming to the present invention and, together with the description, serve to explain the principles of the present invention. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort. In the drawings: Figure 1 This is a schematic diagram illustrating an application scenario of a multi-camera collaborative detection system based on hyperspectral and visible light imaging according to an embodiment of this utility model. Figure 2 This is a module connection diagram of a multi-camera collaborative detection system based on hyperspectral and visible light imaging according to an embodiment of this utility model; Figure 3 This is a schematic diagram of the structure of a light source in an embodiment of this utility model; Figure 4 This is a schematic diagram of another light source in an embodiment of this utility model; Explanation of reference numerals in the attached figures: 1. Material conveyor belt; 2. Light source; 3. Hyperspectral camera; 4. Visible light camera; 5. Encoder; 6. Logic controller; 7. Preparatory sensor; 8. Trigger sensor; 21. Lamp housing; 22. Halogen lamp bead; 211. Heat sink fins; 61. High-speed counter module; 62. Pulse delay timing module; 63. On-delay timing module. Detailed Implementation

[0027] The terminology used in the following embodiments of this utility model is for the purpose of describing particular embodiments only and is not intended to be limiting of the utility model. As used in the specification and appended claims of this utility model, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “or” as used in this utility model refers to and includes any or all possible combinations of one or more of the listed items. Hereinafter, the terms “first” and “second” are used for descriptive purposes only to distinguish technical features and should not be construed as implying relative importance or implicitly indicating the number of indicated technical features. In the description of the embodiments of this utility model, unless otherwise stated, “a plurality” means two or more.

[0028] It should also be noted that, unless otherwise explicitly specified and limited, the terms "setting" and "connection" in the embodiments of this utility model should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this utility model according to the specific circumstances. The embodiments of this utility model are described in detail below.

[0029] This invention provides a multi-camera collaborative detection system based on hyperspectral and visible light imaging. The system uses an integrated logic controller, which is electrically connected to an encoder representing production line displacement, a visible light camera, and a hyperspectral camera. The logic controller utilizes internal hardware circuitry (such as counting and comparison circuits) to process encoder signals and generate precisely delayed camera trigger signals. This hardware architecture solves the problem of difficult-to-achieve precise multi-camera collaboration in existing technologies due to decentralized system control.

[0030] This embodiment constructs a unified, reliable, and high-precision hardware control architecture through this specific hardware composition and connection relationship. It utilizes a displacement reference for synchronization, eliminating errors caused by velocity fluctuations, achieving precise spatial alignment of camera acquisitions, and providing high-quality source images for subsequent multimodal data fusion, thereby improving the overall performance of the detection system. Importantly, the core protection of this invention lies in the aforementioned system hardware configuration and the electrical connections between its components, rather than any software program or control method.

[0031] The following is combined with Figure 1 and 2The system of this embodiment will be described in detail below. The system includes: a material conveyor belt 1, an encoder 5, a hyperspectral camera 3, a visible light camera 4, and a logic controller 6.

[0032] The material conveyor belt 1 is the actuator that carries and moves the material being tested. In the field of food testing, the material conveyor belt 1 is typically made of a flat belt material that meets food-grade safety standards (such as PVC, PU, ​​etc.), and is driven by a drive motor (such as a servo motor or frequency converter motor) through a transmission roller. The operating speed of the material conveyor belt 1 is set to be stable and uniform to ensure that the material being tested can pass through the testing area below at a consistent rate.

[0033] The encoder 5 is a sensor that converts mechanical displacement or angular displacement into electrical signals. In this embodiment, the encoder 5 is preferably a rotary encoder, mounted on the transmission component of the material conveyor belt 1, for example, on the output shaft of the drive motor, the shaft of the transmission roller, or the driven wheel shaft that rolls synchronously with the surface of the material conveyor belt 1. When the material conveyor belt 1 is running, it drives the transmission component to rotate, and the encoder 5 rotates accordingly, continuously generating a series of pulse signals. The number of these pulse signals is proportional to the rotation angle of the transmission component, and thus corresponds precisely to the running displacement of the material conveyor belt 1. For example, after calibration, it can be determined that each pulse signal represents a fixed, minute forward movement of the material conveyor belt 1 (e.g., 0.1 mm). Therefore, the pulse signals generated by the encoder 5 can serve as a high-precision displacement reference, accurately reflecting the positional changes of the detected material on the material conveyor belt 1.

[0034] The hyperspectral camera 3 is positioned above the material conveyor belt 1, with its lens facing downwards and aimed at the detection area of ​​the conveyor belt 1. The hyperspectral camera 3 is an imaging device capable of acquiring image data of an object across tens or even hundreds of consecutive spectral bands. It can simultaneously capture the spatial information (two-dimensional image) and spectral information (third dimension) of the detected material, forming a three-dimensional data cube. The hyperspectral camera 3 has an external trigger interface (e.g., a BNC interface or GPIO pin), which allows precise control of the acquisition time of each frame (or each scan line) of image via externally input level or pulse signals. This external trigger mode is crucial for achieving precise synchronization with external devices.

[0035] The visible light camera 4 can be a standard industrial camera or a digital camera, and its main function is to acquire images of objects in the visible light band (usually red, green, and blue). These images can intuitively reflect the external morphological characteristics of the inspected material, such as color, shape, size, surface texture, and defects. In this system, the visible light camera 4 is also set along the conveying direction of the material conveyor belt 1, and its lens field of view covers the width of the material conveyor belt 1.

[0036] In this embodiment, a preset physical distance exists between the visible light camera 4 and the hyperspectral camera 3. This physical distance refers to the straight-line distance between the imaging center line of the visible light camera 4 and the imaging center line of the hyperspectral camera 3 along the running direction of the material conveyor belt 1. This distance is fixed, measurable, and determined during system installation. Due to this physical distance, without coordinated control, the images captured by the two cameras at the same time do not correspond to the same location of the detected material.

[0037] The logic controller 6 is the control core of the entire collaborative detection system. The logic controller 6 establishes electrical connections with the encoder 5, the hyperspectral camera 3, the visible light camera 4, and the control system (such as a frequency converter or motor drive) of the material conveyor belt 1.

[0038] The function of the logic controller 6 is to receive and process signals and issue control commands. In this embodiment, the logic controller 6 can specifically be a programmable logic controller (PLC), a microcontroller, an MCU, or other controllers or processors; no specific designation is made in this embodiment.

[0039] Taking a PLC controller as an example, this is a stable, reliable, easy-to-program control unit widely used in industrial automation. Specific models available on the market include Siemens' SIMATIC S7-1200 series PLC controllers or Mitsubishi's FX series PLC controllers. These PLC controllers feature high-speed counting input ports, capable of accurately receiving and counting high-frequency pulse signals generated by encoder 5; they also have high-speed pulse output ports or precise timer functions, capable of generating trigger control signals with nanosecond or microsecond precision.

[0040] Next, the workflow of the system in this embodiment will be described in detail.

[0041] After the system is started, the material conveyor belt 1 begins to run, carrying the material to be detected through the detection areas below the visible light camera 4 and the hyperspectral camera 3 in sequence.

[0042] Acquisition of the reference trigger signal: As the material conveyor belt 1 moves, the encoder 5 mounted on its transmission components rotates synchronously, continuously generating pulse signals synchronized with the displacement of the material conveyor belt 1. This series of pulse signals is sent to the dedicated high-speed counting input port of the logic controller 6. The logic controller 6 uses these received pulse signals as the reference trigger signal for the entire system. Here, the reference trigger signal is based on displacement, not time, which ensures that the synchronization accuracy of the system is unaffected by fluctuations in the speed of the material conveyor belt 1.

[0043] Camera control signal generation and timing compensation: The core task of logic controller 6 is to generate a first camera control signal and a second camera control signal with a preset timing difference based on the reference trigger signal, in order to compensate for the preset physical distance between the two cameras. The specific implementation is as follows: First, during the system debugging phase, encoder 5 needs to be calibrated to determine the actual displacement represented by one pulse. Simultaneously, the preset physical distance between visible light camera 4 and hyperspectral camera 3 is accurately measured. Based on these two parameters, logic controller 6 calculates the pulse count value required to compensate for this physical distance. For example, if the preset physical distance is 500 mm, and each pulse represents a displacement of 0.1 mm, then logic controller 6 needs to count 5000 pulses to represent the detected material moving from the position of the first camera to the position of the second camera. This pulse count value is preset in the program of logic controller 6.

[0044] When a portion of the material being detected (e.g., its leading edge) reaches the imaging position of the upstream camera (let's say visible light camera 4), the logic controller 6 can generate a first camera control signal based on an external sensor signal or its own pulse count, and send it to the visible light camera 4. Upon receiving this signal, the visible light camera 4 immediately performs an image acquisition operation.

[0045] At the same moment the first camera control signal is generated, a counter based on the pulses from encoder 5 is started internally by logic controller 6. Logic controller 6 continuously receives and counts the pulse signals generated by encoder 5.

[0046] When the internal counter reaches the preset pulse count value (e.g., 5000), the logic controller 6 determines that the position on the material being detected that was just captured by the visible light camera 4 has now precisely moved below the imaging position of the downstream camera (i.e., the hyperspectral camera 3). At this point, the logic controller 6 immediately generates a second camera control signal and sends it to the hyperspectral camera 3. Upon receiving this signal, the hyperspectral camera 3 also performs an image acquisition operation.

[0047] During this process, there is a delay between the control signals of the first and second cameras. This delay is not a fixed time value, but a fixed pulse count value. This delay, determined by the pulse count value, is the "preset timing difference". Because this timing difference is strictly based on displacement (pulse count), it ensures that the second camera is always triggered at the same position on the material, regardless of the speed of the material conveyor belt 1 or whether the speed changes.

[0048] Spatial alignment is achieved through the control method described above. Although the visible light camera 4 and the hyperspectral camera 3 capture images at different times, the physical location of the material being tested corresponding to the image data acquired by both cameras is consistent due to precise compensation at the trigger time. This achieves spatially aligned acquisition at the same location on the material being tested. Subsequent image processing software can perform pixel-level precise registration and data fusion of these two images from different cameras to obtain the surface morphology features and internal composition information of the material being tested at the same location.

[0049] This invention provides a multi-camera collaborative detection system based on hyperspectral and visible light imaging. By setting up a unified logic controller 6, it utilizes encoder pulses synchronized with the production line displacement as a system reference. Based on this reference and a pulse count value equivalent to the physical distance between the cameras, it generates delayed trigger signals for both cameras. This design successfully solves the problem mentioned in the background art, where the lack of a unified and refined hardware coordination architecture makes it difficult to achieve efficient and accurate collaborative operation when visible light imaging and hyperspectral imaging systems are controlled as independent subsystems.

[0050] The beneficial effects of this embodiment are as follows: First, it achieves high-precision spatial alignment. Because the triggering mechanism is directly linked to the actual displacement of the material, rather than relying on speed estimation and time delay, it fundamentally eliminates synchronization errors caused by production line speed fluctuations, ensuring that the same position of the material can be imaged under any operating condition.

[0051] Secondly, a unified and reliable hardware control architecture was constructed. An industrial-grade PLC was adopted as the logic controller, integrating the production line displacement sensors and all camera trigger controls into a single hardware platform. This replaced the previous decentralized control method, which might have relied on host computer software, and greatly improved the system's stability and response speed.

[0052] Third, it improves the overall performance and scalability of the detection system. Precise spatial alignment enables subsequent multimodal data fusion (e.g., correlation analysis of surface defects detected by visible light images with component anomalies detected by hyperspectral images), thereby significantly improving detection accuracy and dimensionality. The architecture is also easily expandable; adding more cameras in the future only requires adding corresponding delay-triggered logic to logic controller 6, demonstrating good flexibility and scalability.

[0053] It is particularly important to emphasize that the innovation of the technical solution claimed in this utility model lies entirely in the hardware configuration of the system and the defined electrical connections between the various hardware components, and not in any improvement to the control method or software program. The scope of protection of this utility model is limited to the system structure as defined in the claims.

[0054] Therefore, it should be understood that what this utility model claims protection for is this hardware combination with a specific connection relationship and cooperative function, and not for improvements to the specific control method or software program to achieve this function. This utility model also does not claim protection for related control methods or software programs.

[0055] In this embodiment, although the logic controller 6 can be implemented using programmable devices such as PLCs in practice, within the protection context of this utility model, the logic controller 6 should be understood as a hardware entity that implements specific functions. All its control functions are implemented through its inherent, dedicated electronic circuits and their specific interconnection configuration. Specifically, the logic controller 6 integrates hardware functional modules such as high-speed pulse counting circuits, numerical comparison circuits, logic gate circuits, and delay-triggered output circuits.

[0056] From the perspective of signal flow and physical implementation, the pulse signal output by encoder 5 is connected to a dedicated high-speed counting input port on logic controller 6 via physical wires. This signal is directly processed by the internal hardware counting circuit. The count value of this counting circuit is provided to the hardware numerical comparison circuit in real time via internal circuit connections (rather than via software bus calls). When the numerical comparison circuit detects that the count value is equal to a preset threshold representing the physical distance between the cameras (this threshold itself can be considered an adjustable parameter of the hardware circuit), the physical state of the comparison circuit flips, directly or through logic gates, triggering the delayed trigger output circuits connected to different physical output ports. Finally, these two output ports transmit control signals to visible light camera 4 and hyperspectral camera 3 via physical lines, respectively.

[0057] Therefore, from receiving a pulse signal from encoder 5 to finally outputting two camera control signals with a preset timing difference, the entire process is a deterministic electrical signal processing path completed at the hardware level. The implementation of this process does not depend on the sequential execution of specific software programs, algorithms, or instruction sequences running on a general-purpose processor. Even though the logic controller 6 requires parameter configuration with the aid of external tools during initial setup, its function as a component of this invention is entirely determined by its fixed hardware logic structure.

[0058] In summary, the contribution of this utility model to the prior art lies in providing a novel hardware system architecture. This architecture inherently and necessarily realizes spatial alignment acquisition control of different cameras by physically combining and electrically interconnecting the encoder 5, the logic controller 6 with a specific internal circuit structure, the visible light camera 4, and the hyperspectral camera 3.

[0059] In some embodiments, the logic controller 6 integrates a high-speed counter module 61. This high-speed counter module 61 is a dedicated hardware circuit unit within the logic controller 6, not a software program.

[0060] The high-speed counter module 61 has dedicated physical input pins that directly connect to the signal output of the encoder 5. Its internal circuitry is designed for hardware-level accumulation and counting of high-frequency level-change signals (i.e., pulse signals), achieving a counting speed at the megahertz (MHz) level, far exceeding the processing capabilities of ordinary software cyclic scanning. In this system, the high-speed counter module 61 receives all pulse signals from the encoder 5 and accumulates them continuously. When a preset timing difference needs to be achieved, for example, after triggering the first camera (such as visible light camera 4), other hardware circuits within the logic controller 6 use the current count value as a reference and calculate a target count value based on a preset number of pulses (e.g., 5000) corresponding to the physical distance between the cameras. When the hardware count value of the high-speed counter module 61 reaches this target count value, its internal hardware comparison circuit directly outputs a trigger signal to trigger the second camera (such as hyperspectral camera 3).

[0061] This purely hardware-based counting and comparison method greatly improves the accuracy of counting and the real-time performance of the response, thus providing a reliable hardware structure guarantee for achieving precise preset timing differences.

[0062] In some embodiments, the logic controller 6 integrates a pulse delay timing module 62. This pulse delay timing module 62 is also a hardware circuit unit within the logic controller 6.

[0063] The function of the pulse delay timing module 62 is to start precise delay timing after receiving an external or internal start signal, and output a signal after the timing ends.

[0064] In this system, the pulse delay timing module 62 can be used to implement a preset timing difference. One implementation is as follows: when the leading edge of the detected material triggers the first camera to take a picture, the arithmetic circuit in the logic controller 6 can quickly calculate a required time delay value based on the pulse frequency fed back by the current encoder 5 (representing the speed of the material conveyor belt 1) and the known physical distance between the cameras. This time delay value is loaded into the hardware register of the pulse delay timing module 62, and the pulse delay timing module 62 is started. The pulse delay timing module 62 then begins to count down using its internal high-precision clock source. When the countdown time reaches the set value, the hardware circuit of the pulse delay timing module 62 will directly output a trigger signal to the second camera.

[0065] This method implements time-based delay through a hardware timer, which also constitutes a specific hardware structure for achieving camera synchronization.

[0066] In this embodiment, by setting a pulse delay timing module 62, the system can calculate a precise time delay value based on the real-time speed after the first camera is triggered, and then hand it over to the hardware timing module for execution. This method also avoids the uncertainty caused by software delay, ensuring the real-time performance and accuracy of the trigger signal generation. It is a flexible and effective hardware structure for realizing camera collaborative control.

[0067] In some embodiments, to ensure imaging quality, the system further includes a light source 2. The light source 2 is physically mounted above the material conveyor belt 1, and its position is carefully designed to ensure that the light emitted by it can uniformly and stably illuminate the detection area covered by the camera's field of view.

[0068] To achieve coordinated control of the light source 2, the logic controller 6 is also specially equipped with a control output terminal for the light source 2. This output terminal is typically one or more relay output points or transistor output points. The control input terminal of the light source 2 (e.g., the control line of its power switch) is connected to the control output terminal of the light source 2 of the logic controller 6 via physical wires.

[0069] With this configuration, the logic controller 6 can directly control the on and off of the light source 2 by outputting high and low level signals through this output terminal. The technical advantage is that it allows the illumination of the light source 2 to be synchronized with the camera's acquisition action; for example, the light source 2 can be turned on only for a short period of time before or during camera exposure, thereby significantly saving energy and extending the lifespan of the light source 2.

[0070] In some embodiments, to further enhance the intelligence and stability of the system, the logic controller 6 is also provided with a conveyor belt signal input terminal. This input terminal is a physical digital input point used to receive a stable operating status signal from the material conveyor belt 1 drive system (e.g., a frequency converter). This signal indicates that the material conveyor belt 1 has started and reached a stable operating speed.

[0071] Meanwhile, the logic controller 6 also integrates a power-on delay timing module 63. The function of the power-on delay timing module 63 is to delay the output signal by a preset time after receiving the start signal. Its operation is as follows: when the conveyor belt signal input receives a stable operating status signal, this signal serves as a trigger signal to activate the power-on delay timing module. The power-on delay timing module 63 begins timing; the delay time can be preset (e.g., 500 milliseconds) to allow the light source 2 (especially the halogen lamp) to reach stable brightness and color temperature. After the delay time, the power-on delay timing module 63 outputs a working start signal. This signal is transmitted through the internal circuitry of the logic controller 6 to the corresponding output terminals controlling the visible light camera 4, the hyperspectral camera 3, and the light source 2, allowing them to begin normal operation.

[0072] The beneficial effect of this technical solution is that it ensures that the entire detection system only starts when both the production line and the light source 2 are in the most stable state, avoiding the problem of invalid or inaccurate data acquisition caused by speed fluctuations or instability of the light source 2 during the startup phase, thus improving the overall reliability of the system.

[0073] In some embodiments, the system further includes a pre-sensor 7 and a trigger sensor 8. These are two independent physical sensors, such as fiber optic sensors or photoelectric switches. The pre-sensor 7 and trigger sensor 8 are arranged sequentially and at intervals upstream of the camera detection area along the material conveying direction of the material conveyor belt 1. The logic controller 6 is provided with corresponding input terminals, which are electrically connected to the signal output terminals of the pre-sensor 7 and the trigger sensor 8, respectively.

[0074] Its operation is as follows: When the material to be detected passes through the pre-sensor 7, the logic controller 6 receives a signal and can perform some preparatory actions in advance, such as turning on the light source 2 for preheating. Subsequently, when the leading edge of the material reaches the position of the trigger sensor 8, the logic controller 6 receives a trigger signal, uses this position as the starting zero point for data acquisition, and begins to use pulse counting based on the encoder 5 to precisely control subsequent camera imaging. Its technical advantage lies in providing the system with a precise event triggering mechanism based on discrete materials, enabling the system not only to scan continuously but also to efficiently handle individual materials that are not continuously distributed on the conveyor belt. Data acquisition is only initiated when the material arrives, making it more flexible and efficient.

[0075] In this embodiment, the pre-sensor 7 and the trigger sensor 8 are essentially the same device in hardware (e.g., both can be Keyence PZ-G series products, such as PZ-G41N). Their difference lies entirely in their physical installation positions on the material conveyor belt 1, thus assuming different logical roles in the system. Their core function is to utilize photoelectric principles to non-contactly detect the presence or absence of the measured material and output a digital switching signal.

[0076] The pre-sensor 7 and the trigger sensor 8 are highly integrated into a robust, sealed housing. They primarily consist of an infrared or red visible light-emitting diode (LED) as a transmitter, a phototransistor or photodiode as a receiver, a precision optical lens system, and circuitry for signal amplification, processing, and output. The transmitter and receiver are mounted side-by-side at a specific angle inside the sensor.

[0077] When the sensor is powered on, its internal LED emits a beam of light that is focused by a lens. When no material is being detected in the detection area, the light beam is directed forward. Since there is no object to reflect the light, the receiver does not receive enough light, and its internal phototransistor is in the off state. At this time, the sensor's output circuit outputs a default level (e.g., a high level, called the "normally open" state).

[0078] When the material being detected (such as a cardboard box, parts, or food bag) enters the effective range of the detection beam, its surface diffusely reflects a portion of the beam back to the sensor. Upon receiving this reflected light, the receiver activates its internal phototransistor, which conducts due to the illumination. The resulting current, after being processed by the internal amplification circuit, drives the output circuit to flip, outputting a level opposite to the default state (e.g., a low level). This level transition is the hardware signal indicating that an object has been detected.

[0079] The pre-detection sensor 7 is physically installed at the very beginning of the detection line. When the material passes through it for the first time, the level transition signal it outputs is captured by the logic controller 6 as a pre-detection signal that "the material is about to enter the detection area".

[0080] The trigger sensor 8 is physically mounted at a precise, fixed reference position close to the camera's field of view. When the material's leading edge reaches the sensor, its output level transition signal serves as a precise trigger signal indicating that "the material has reached the acquisition start point," which is used to start the encoder 5's pulse counting or directly trigger the first camera.

[0081] In this embodiment, the encoder 5 is an electromechanical conversion hardware device that converts the mechanical angular displacement of the physical world into digital electrical signals. Its core function is to provide the logic controller 6 with a high-precision pulse signal source that is strictly proportional to the displacement of the material conveyor belt 1, based on purely physical and optical principles.

[0082] Hardware Composition and Working Principle: In this system, an incremental photoelectric encoder 5 is typically used. Its core internal hardware structure mainly includes: a glass or metal "grating disk" with a central shaft hole, a light-emitting diode (LED) as the light source 2, a receiver array (photoelectric integrated circuit) containing multiple phototransistors, and shaping and output circuits.

[0083] Code Wheel: This is a thin disc with hundreds to thousands of opaque lines evenly distributed on its surface through etching or printing processes, with light-transmitting slits between the lines. The code wheel is rigidly fixed to the rotating shaft of encoder 5, which is mechanically coaxially connected to the transmission components (such as roller shafts) of material conveyor belt 1.

[0084] Optoelectronic transceiver assembly: On one side of the grating disk is a fixed, highly stable light-emitting diode (LED), whose emitted light (usually infrared light) is directed parallel to the grating disk. On the other side of the grating disk, directly opposite light source 2, is a fixed, highly integrated optoelectronic receiver.

[0085] Signal generation process: When the material conveyor belt 1 drives the transmission components to rotate, the grating disk rotates accordingly. The grating disk's scribe lines periodically block and allow the light emitted by the LEDs to pass through. Therefore, the photodetector opposite receives an alternating light signal of varying brightness. The phototransistor inside the photodetector directly converts this light signal variation into a change in current or voltage. This analog changing electrical signal is then sent to the hardware "shaping circuit" (usually a Schmitt trigger circuit) inside the encoder 5. The function of this circuit is to convert the approximately sinusoidal, potentially interference-laden, analog signal into a steep, jitter-free standard square wave digital pulse signal.

[0086] Signal Output: The shaped square wave pulse signal is ultimately driven by a hardware output circuit (such as an NPN / PNP open collector or push-pull output circuit) to the output pin, and then connected to the high-speed counting module of the logic controller 6 via a physical cable. For every tiny fixed distance the material conveyor belt 1 moves, the encoder 5 outputs a complete square wave pulse.

[0087] In some embodiments, the light source 2 is a custom halogen light source. This custom halogen light source is a hardware assembly, such as... Figure 3 and Figure 4As shown, its internal structure includes at least three core components: lamp housing 21, halogen lamp beads 22, and voltage regulator module (not shown).

[0088] The halogen lamp bead 22 is the actual light-emitting element, capable of emitting light with continuous spectrum characteristics. The halogen lamp bead 22 is installed and fixed inside the lamp housing 21. The voltage regulator module is an electronic power device capable of converting unstable input electrical energy into highly stable output electrical energy in terms of voltage or current; for example, it can be a power module based on switching power supply (SMPS) technology or linear regulation technology. This voltage regulator module is structurally electrically connected to the halogen lamp bead 22, specifically providing it with the electrical energy required for operation.

[0089] This embodiment clearly defines the internal hardware structure of the light source 2, particularly by introducing the key component of the voltage regulator module, ensuring extremely high stability of the light intensity output by the light source 2. The luminous flux of the halogen lamp 22 is extremely sensitive to fluctuations in the supply voltage; any tiny voltage fluctuation will cause drastic changes in light intensity, thus forming bright and dark stripe noise in the spectral image. The voltage regulator module effectively filters out interference caused by mains voltage fluctuations or load changes, providing a constant driving current or voltage to the halogen lamp 22, thereby ensuring that the light source 2 can output stable, flicker-free illumination throughout its operation. This is a prerequisite hardware guarantee for obtaining high-quality, high signal-to-noise ratio spectral data.

[0090] In some embodiments, the number of halogen lamp beads 22 is multiple, and these halogen lamp beads 22 are arranged in a specific geometric shape inside the lamp housing 21, such as a linear array or a ring array. A linear array refers to arranging multiple halogen lamp beads 22 in a close or equally spaced manner along a straight line.

[0091] During installation, this straight line direction is typically parallel to the scanning line direction of the hyperspectral camera 3 (especially a line scan camera). A ring array refers to arranging multiple halogen lamps 22 along a circumference to form a ring-shaped light source 2.

[0092] This embodiment achieves uniformity and customization of the illumination area to match specific camera field of view and detection requirements. By employing a linear array arrangement, a narrow and high-intensity linear illumination area can be formed above the detection area of ​​the material conveyor belt 1. This illumination area perfectly matches the slit field of view of the line-scan hyperspectral camera 3, enabling the light energy to be highly concentrated in the effective imaging area, improving light energy utilization and ensuring uniform illumination at all points along the scanning line.

[0093] Using a ring array can create a regularly shaped area of ​​illumination, suitable for area scan cameras or situations requiring diffuse illumination from multiple angles to eliminate shadows. This specific hardware arrangement structure solves the problems of uneven illumination and bright center with dark periphery caused by a single point light source 2 or randomly arranged light sources 2, providing a structural guarantee for obtaining high-quality data with uniform brightness across the entire image.

[0094] In some embodiments, the lamp housing 21 itself is a metal heat sink, and heat dissipation fins 211 are integrally formed or additionally provided on the outer surface of the metal heat sink. The lamp housing 21 is typically made of a metal with high thermal conductivity, such as 6063 series aluminum alloy profiles.

[0095] The heat dissipation fins 211 refer to a series of sheet-like or columnar structures extending from the surface of the outer casing, the purpose of which is to significantly increase the contact area between the outer casing and the surrounding air.

[0096] The design of the lamp housing 21 in this embodiment enhances the passive heat dissipation capability of the light source 2, thereby ensuring the stability and long lifespan of the light source 2. Halogen LEDs 22 generate a large amount of waste heat when emitting light. If this heat cannot be dissipated in time, it will lead to excessively high LED temperatures, resulting in accelerated light decay, color temperature drift, and even burnout. By employing a metal heat sink with excellent thermal conductivity, the heat generated by the halogen LEDs 22 can be rapidly conducted to the entire housing; while numerous heat dissipation fins 211 on the outer surface efficiently dissipate this heat to the surrounding environment through air convection and radiation. This hardware structure design ensures that the halogen LEDs 22 can operate at a reasonable and constant temperature, thereby guaranteeing the long-term stability of their light output parameters, significantly extending the overall lifespan of the light source 2, and reducing maintenance costs.

[0097] In some embodiments, the color temperature of the custom halogen light source 2 is selected and controlled within the range of 2800K to 3200K.

[0098] Color temperature, in this context, refers to the temperature at which an ideal blackbody, when heated to a certain temperature, emits light of the same color as a given light source (source 2). The color temperature range of 2800K to 3200K falls under the category of warm white light. Light sources in this range exhibit a balanced spectral energy distribution across the visible and near-infrared bands, making them highly suitable as illumination sources for hyperspectral imaging systems.

[0099] In this embodiment, the core performance parameters of the hyperspectral camera are specifically defined. The spectral range of the hyperspectral camera used is 400nm-2500nm, and its spectral resolution is 1nm-10nm.

[0100] The spectral range refers to the wavelength range of electromagnetic waves that the camera can detect. The 400nm-2500nm range covers the visible light (approximately 400-780nm), near-infrared (approximately 780-1100nm), and short-wave infrared (approximately 1100-2500nm) regions. Spectral resolution refers to the smallest wavelength interval that the camera can distinguish. These hardware performance parameters of the hyperspectral camera are included in the claims because this specific hardware selection directly determines the detection capability and application breadth of the system. Its technical advantages lie in the fact that covering such a wide spectral range allows the system not only to identify color differences but also to detect chemical components such as moisture, proteins, fats, and sugars, as well as different types of plastics and minerals, making its applications extremely broad.

[0101] A spectral resolution of 1nm-10nm ensures sufficient resolution while also guaranteeing a high signal-to-noise ratio and data acquisition speed, representing an optimal balance for industrial online inspection applications. Therefore, this hardware choice is a crucial structural foundation for realizing the powerful detection capabilities of this invention.

[0102] In some embodiments, the spectral range of the hyperspectral camera 3 is preferably 900 nm to 1700 nm, and the spectral resolution is 8 nm.

[0103] Spectral range is defined as the wavelength range of electromagnetic waves that the camera's image sensor can effectively detect. 900nm to 1700nm falls within the near-infrared (NIR) or short-wave infrared (SWIR) band. Spectral resolution is defined as the smallest wavelength interval that a spectrometer can resolve; a smaller value indicates finer resolution. An 8nm spectral resolution means that within the total bandwidth of 800nm ​​(1700nm-900nm), the hyperspectral camera 3 can divide approximately 100 (800 / 8) independent, continuous spectral channels to acquire data. The beneficial effect of this technical solution is that it endows the entire system with the ability to detect specific internal quality defects.

[0104] Scientific research and practical applications show that the absorption peaks of molecular vibrations of the chemical components inside organic matter (such as chili peppers), such as water, protein, and specific chemical substances produced during the mold growth process, are mostly located in the spectral range of 900nm to 1700nm.

[0105] Therefore, the selection of a hyperspectral camera 3 within this spectral range enables the system to "see through" the material surface and directly detect characteristic spectral information reflecting internal chemical changes such as mold growth. The 8nm spectral resolution ensures the capture of the fine shapes of these characteristic absorption peaks while avoiding the data explosion and processing speed reduction problems associated with excessively high resolutions (such as 1-2nm). This achieves a technical balance between detection accuracy and processing efficiency, making the system practically feasible in industrial high-speed online detection applications.

[0106] The workflow of the system in this embodiment is a highly automated, closed-loop process with precise hardware logic coordination. The following will describe in detail the process from power-on startup to the completion of a full test, following the time sequence and signal flow.

[0107] Step 1: Power on and initialize the system.

[0108] Once the main power supply to the entire testing line is turned on, all hardware components, including the logic controller, encoder, hyperspectral camera, visible light camera, custom halogen light source, pre-sensor, and trigger sensor, receive power. At this time, the logic controller performs an internal hardware self-test and enters standby mode. The voltage regulator module inside the custom halogen light source begins to operate, providing stable power to the halogen LEDs, but the light source control output is currently off, and the halogen LEDs do not emit light.

[0109] Step 2: Wait for the production line to run stably.

[0110] The operator starts the material conveyor belt. After startup, the conveyor belt drive system (such as a frequency converter) outputs a "stable operation status signal." This signal is sent via physical lines to the dedicated "conveyor belt signal input terminal" on the logic controller. Upon receiving this signal, the "on-delay timing module" hardware circuit inside the logic controller is triggered and begins a hardware delay timer for a preset period (e.g., 500 milliseconds).

[0111] Step 3: The system enters the preparation state.

[0112] After the delay of the turn-on delay timing module ends, the module's hardware circuit outputs a "work-on" signal. This signal, through the internal circuitry of the logic controller, simultaneously affects three parts: It enables the trigger input ports of visible light cameras and hyperspectral cameras, putting them into a state ready to receive trigger signals.

[0113] It sends an on signal to the customized halogen light source through the "light source control output terminal" to illuminate the light source. Due to the physical characteristics of halogen lamps, a short warm-up time is required to achieve a stable color temperature and light intensity of 2800K to 3200K, and the aforementioned delay timing provides just that.

[0114] It resets and activates the "high-speed counter module" connected to the encoder, preparing it to count the encoder's pulses. At this point, the system has completed all preparations and enters a state awaiting the material to be detected. The detection area above the material conveyor belt is covered by a uniform linear or planar illumination area formed by a linear or circular array of halogen lamps.

[0115] Step 4: Material entry and preparation triggering.

[0116] The material being detected moves forward with the material conveyor belt. When the leading edge of the material first reaches and blocks or reflects the beam of the "preparatory sensor," the state of the preparatory sensor hardware circuit flips, outputting a level transition signal to the logic controller. Upon receiving this preparatory signal, the logic controller can perform some preprocessing actions, such as confirming whether the system components are ready, but at this time, the core image acquisition is not initiated.

[0117] Step 5: Start precise positioning and synchronous data acquisition.

[0118] The material continues to move forward, and when its leading edge precisely reaches the installation position of the "trigger sensor," the hardware state of the trigger sensor also flips, outputting a precise trigger signal to the logic controller. The logic controller defines the arrival time of this signal as the "zero point" of the data acquisition process. At this moment, the logic controller immediately executes the following two parallel hardware actions: For the upstream camera (assuming it's a visible light camera): the logic controller immediately sends a first camera control signal (a pulse signal) to the visible light camera through its corresponding control signal output. Upon receiving this signal, the visible light camera completes an image acquisition of the material at its current location.

[0119] For the downstream camera (assuming it's a hyperspectral camera): At the same moment the first camera control signal is output, the hardware circuitry within the logic controller locks the current count value of the "high-speed counter module" to a reference value. The logic controller adds the locked reference value to a pre-configured number of pulses representing the preset physical distance between the visible light camera and the hyperspectral camera (this number is calculated by dividing the "physical distance" by the "unit pulse displacement after encoder calibration") to obtain a target count value.

[0120] Step 6: Implement displacement-based delayed triggering and spatial alignment.

[0121] The material conveyor belt operates continuously, its transmission components driving the encoder to rotate constantly. The encoder hardware continuously generates pulse signals synchronized with the displacement, which are counted in real time by the "high-speed counter module." The hardware comparison circuit within the logic controller continuously compares the current real-time count value of the high-speed counter module with the "target count value" calculated in the previous step. Only when these two values ​​are completely equal does it mean that the material position captured by the visible light camera has now precisely moved to directly below the hyperspectral camera. At this point, the state of the hardware comparison circuit immediately flips, directly triggering the logic controller to send a second camera control signal to the hyperspectral camera through another control signal output terminal. Upon receiving the signal, the hyperspectral camera immediately completes an image acquisition. The hyperspectral camera has a spectral range of 400nm-2500nm and a spectral resolution of 1nm-10nm, ensuring rich and accurate image information.

[0122] Step 7: Loop detection and system shutdown.

[0123] This concludes a complete spatial alignment acquisition process. The acquired visible light and hyperspectral images are transmitted by their respective cameras to the subsequent image processing system for analysis. The logic controller then returns to the state of waiting for the next trigger sensor signal, preparing to detect the next material, thus achieving continuous and efficient online detection. When the entire batch of materials has been detected and the operator stops the material conveyor belt, the "stable operating status signal" from the drive system disappears, the logic controller automatically shuts off the light source, and returns the camera and other components to standby mode, ending the entire system process.

[0124] The entire workflow is driven entirely by signal interaction between hardware components and the internal hardware circuitry of the logic controller, without relying on real-time intervention from the host computer software, thus ensuring extremely high synchronization accuracy, response speed, and system stability.

[0125] The above description is merely a preferred embodiment of this utility model, but the protection scope of this utility model is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this utility model should be included within the protection scope of this utility model. Therefore, the protection scope of this utility model should be determined by the scope of the claims.

Claims

1. A multi-camera collaborative detection system based on hyperspectral and visible light imaging, characterized in that, include: Material conveyor belts are used to transport materials to be tested. An encoder, installed on the transmission component of the material conveyor belt, is used to generate a pulse signal synchronized with the running displacement of the material conveyor belt; A hyperspectral camera is set along the conveying direction of the material conveyor belt to acquire hyperspectral images of the material being tested. A visible light camera is set along the conveying direction of the material conveyor belt to acquire visible light images of the material being detected. There is a preset physical distance between the visible light camera and the hyperspectral camera. The logic controller is connected to the encoder, the hyperspectral camera, the visible light camera, and the material conveyor belt, respectively. The logic controller is used to receive the pulse signal generated by the encoder as a reference trigger signal, and based on the reference trigger signal, generate a first camera control signal and a second camera control signal with a preset timing difference, and send the first camera control signal to the visible light camera and the second camera control signal to the hyperspectral camera to compensate for the preset physical distance, thereby controlling the image acquisition of the visible light camera and the hyperspectral camera to achieve spatial alignment at the same position of the material being detected.

2. The system according to claim 1, characterized in that, The logic controller is equipped with a high-speed counter module, which is used to count the pulse signals received from the encoder in order to achieve the preset timing difference.

3. The system according to claim 1, characterized in that, The logic controller is equipped with a pulse delay timing module, which is used to perform a delay timing after receiving a pulse signal from the encoder in order to achieve the preset timing difference.

4. The system according to any one of claims 1-3, characterized in that, It also includes a light source, which is positioned above the material conveyor belt; the logic controller is further provided with a light source control output terminal, and the control input terminal of the light source is connected to the light source control output terminal.

5. The system according to claim 4, characterized in that, The logic controller is also provided with a conveyor belt signal input terminal and a turn-on delay timing module. The conveyor belt signal input terminal is used to receive the stable operating status signal of the material conveyor belt. The turn-on delay timing module is used to delay sending a working start signal to the visible light camera, the hyperspectral camera and the light source after receiving the stable operating status signal at the conveyor belt signal input terminal.

6. The system according to claim 4, characterized in that, The light source is a custom halogen light source, which includes a lamp housing, halogen lamp beads, and a voltage regulator module; the halogen lamp beads are disposed inside the lamp housing, and the voltage regulator module is used to supply power to the halogen lamp beads.

7. The system according to claim 6, characterized in that, Multiple halogen lamp beads are arranged in a linear or circular array to form a uniform linear or planar lighting area above the material conveyor belt.

8. The system according to claim 6, characterized in that, The lamp housing is a metal heat dissipation shell, and the outer surface of the lamp housing is provided with heat dissipation fins. The color temperature of the customized halogen light source is in the range of 2800K to 3200K.

9. The system according to claim 1, characterized in that, It also includes a pre-sensor and a trigger sensor; the pre-sensor and the trigger sensor are arranged sequentially at intervals along the material conveying direction of the material conveyor belt; the logic controller is connected to the pre-sensor and the trigger sensor respectively.

10. The system according to claim 1, characterized in that, The hyperspectral camera has a spectral range of 400nm-2500nm and a spectral resolution of 1nm-10nm.

Images