Anti-severe interference imaging system and equipment

Abstract

The invention relates to the field of detection imaging, in particular to an anti-severe interference imaging system and device. The system comprises an infrared imaging module, a laser radar module, a single-photon radar module and a data processing unit. The data processing unit is used for acquiring first detection data, second detection data and third detection data; according to the environment parameters of the target environment, the confidence coefficient weight of the first detection data and the fusion weight of the second detection data and the third detection data are adjusted; generating a first contour feature based on the first detection data, and filtering the second detection data and the third detection data according to the first contour feature and a preset error threshold to obtain target second detection data and target third detection data; the preset error threshold value is determined according to the confidence coefficient weight; and based on the fusion weight, fusing the target second detection data and the target third detection data to generate a target image. And the high-robustness target sensing capability is realized in a complex interference environment.

Description

Technical Field

[0001] This application relates to the field of detection imaging, and more particularly to an imaging system and device resistant to severe interference. Background Technology

[0002] LiDAR (Light Detection and Ranging) is a radar monitoring technology that achieves high-precision ranging, 3D modeling, and dynamic target perception by emitting laser pulse beams and receiving the echo signals reflected from targets. Its core principle utilizes the coherence, directionality, and high energy density of laser light to acquire target information through Time-of-Flight (ToF) or the Doppler effect. LiDAR is widely used in autonomous driving, military reconnaissance, and security monitoring.

[0003] For example, patent application CN112731443A discloses a three-dimensional imaging system and method for fusing single-photon lidar with short-wave infrared images. The system includes a laser source, a single-photon detector, an optical path system, a signal control unit, and a short-wave infrared camera. The laser source includes a laser controller connected to the single-photon detector. The optical path system is a coaxial transmit / receive optical path, including a transmitting optical path and a receiving optical path. The signal control unit is connected to the single-photon detector and the laser source and is used to emit control signals.

[0004] For example, patent application CN119618194A discloses a sensing device and method based on multimodal sensor integration and synchronization. The processing unit and power supply unit are located inside the main support, with the processing unit mounted above the power supply unit. A display unit is located at one end of the main support, and a sensor unit is mounted on the top of the main support. A sensor synchronization module is located on one side of the sensor unit. The power supply unit is electrically connected to the processing unit, and the processing unit is electrically connected to its sensor unit, display unit, and sensor synchronization module. A hardware trigger circuit is designed to simultaneously trigger the LiDAR, visible light camera, and infrared camera to collect data. Then, a memory-sharing method is used to align the timestamps of each frame of data from the three cameras. Through the integration of the LiDAR, visible light camera, and infrared camera, and a high-precision time synchronization method, comprehensive and high-quality data input is provided for the SLAM system.

[0005] However, harsh environments are subject to various interferences, and the performance of various detection modules is easily affected by these interferences, resulting in insufficient adaptability and reliability of existing imaging technologies in harsh environments. Summary of the Invention

[0006] The main objective of this application is to provide an imaging system and device resistant to severe interference. To solve the aforementioned technical problems, this application specifically adopts the following technical solution: The first aspect of this application is to provide an imaging system resistant to severe interference, the system comprising an infrared imaging module, a lidar module, a single-photon radar module, and a data processing unit; The data processing unit is used for: Acquire the first detection data collected by the infrared imaging module, the second detection data collected by the lidar module, and the third detection data collected by the single-photon radar module; Based on the environmental parameters of the target environment, the weights among the first detection data, the second detection data, and the third detection data are adjusted; wherein, the weights include the confidence weight of the first detection data and the fusion weight of the second and third detection data; A first contour feature is generated based on the first detection data, and the second and third detection data are filtered according to the first contour feature and a preset error threshold to obtain target second detection data and target third detection data; wherein, the preset error threshold is determined according to the confidence weight; Based on the fusion weights, the second and third target detection data are fused to generate a target image.

[0007] A second aspect of this application is to provide an imaging device resistant to severe interference, the device being applied to the imaging system resistant to severe interference provided in any embodiment of this application, the device comprising: The infrared imaging module is used to collect the first detection data corresponding to the target environment; The lidar module is used to collect secondary detection data corresponding to the target environment; The single-photon radar module is used to collect third-party detection data corresponding to the target environment.

[0008] Beneficial effects: This application provides an imaging system and device resistant to severe interference. It acquires heterogeneous and complementary detection data through triple detection of infrared imaging, lidar and single-photon radar, and proposes a multi-dimensional and highly collaborative hierarchical filtering mechanism based on a multi-modal module. Basic interference filtering is achieved through physical and algorithmic collaboration in the early stage of acquisition, and a closed-loop filtering with dynamic weights is constructed in the data processing stage. This achieves highly robust target perception capability in complex interference environments.

[0009] Firstly, basic filtering is performed during the initial data acquisition phase. The single-photon radar module locks onto the core time window of the target echo through main peak detection, receiving only photon signals within that time period to suppress scattering noise and environmental noise interference. Its polarization beam splitter separates the emitted laser and echo signals through waveplate angle adjustment, further enhancing anti-interference capabilities. The lidar module receives corresponding point cloud data through a shared time window, eliminating a large number of useless background or noise points. Simultaneously, both the lidar and infrared imaging modules integrate optical phase change materials (such as vanadium dioxide) to dynamically reduce the transmittance of the corresponding band when the incident signal intensity is abnormal, suppressing strong light interference. In addition, the timing controller synchronizes the trigger signals of the three types of sensors to ensure data alignment in the spatiotemporal dimensions, laying the foundation for subsequent fusion processing.

[0010] Secondly, dynamic weighted filtering is implemented during the processing stage. The system dynamically adjusts the confidence weight and fusion weight of the detection data based on environmental parameters to adapt to the different adaptability of different modules in various harsh environments. Using the infrared data-generated contour as a benchmark, a dynamic error threshold corresponding to the confidence weight is set to verify and filter laser and single-photon point clouds. Furthermore, a reverse verification and compensation mechanism is introduced: for abnormally missing continuous contours, the error threshold is temporarily relaxed, and suspicious data is re-verified and a reasonable contour is completed. This closed-loop strategy of "filtering-verification-compensation" retains key target information while avoiding the introduction of interfering data, ultimately achieving reliable target reconstruction in complex interference environments. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. The elements or parts in the drawings are not necessarily drawn to scale. Obviously, the drawings described below are some embodiments of this application; for those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0012] Figure 1 This is a schematic diagram of an imaging system resistant to severe interference provided in an embodiment of this application; Figure 2 This is a structural schematic diagram of an imaging system resistant to severe interference provided in an embodiment of this application; Figure 3 This is a flowchart illustrating the data processing process of a data processing unit according to an embodiment of this application. Figure 4 This is a flowchart illustrating the data processing process of another data processing unit provided in this application embodiment. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0015] In this document, suffixes such as “module,” “part,” or “unit” used to denote elements are used only for illustrative purposes and have no specific meaning in themselves. Therefore, “module,” “part,” or “unit” may be used interchangeably.

[0016] In this document, the terms "upper," "lower," "inner," "outer," "front," "rear," "one end," and "the other end," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0017] In this document, unless otherwise explicitly specified and limited, the terms "installed," "equipped with," and "connected," etc., 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, 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 application based on the specific circumstances.

[0018] In this document, "and / or" includes any and all combinations of one or more of the listed related items.

[0019] In this article, "multiple" means two or more, that is, it includes two, three, four, five, etc.

[0020] In this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0021] In this paper, adverse environments refer to external environmental conditions that significantly negatively impact the performance of the detection system. These mainly include: severe weather interference, light pollution interference, electromagnetic interference, and thermal radiation interference. Severe weather environments refer to high-scattering environments such as rain, snow, fog, and dust storms, which enhance atmospheric scattering and absorption effects. Particles in these environments cause Mie scattering of the laser signal, resulting in signal attenuation, increased noise, and degraded target features. The extent of this impact is closely related to particle size, concentration, and atmospheric visibility. Light pollution interference environments include direct sunlight and artificial light sources such as urban lighting, whose energy covers the detection band, causing sensor saturation or introducing false echoes. Electromagnetic interference environments consist of electromagnetic radiation generated by radar, motors, and wireless communication equipment, which can lead to system signal distortion or a decrease in the signal-to-noise ratio. Thermal radiation interference environments refer to scenarios with strong infrared radiation sources such as thermal decoys and high-temperature backgrounds, which may mask the thermal characteristics of the true target.

[0022] In this paper, the target environment refers to the specific external environment that the current detection system needs to detect, and its environmental state directly affects the detection performance of each sensor. Environmental parameters of the target environment refer to physical quantities or characteristic information used to characterize the external environment and its impact on the quality of detection data. These parameters include at least one or more of the following: atmospheric visibility, light intensity, background thermal radiation level, electromagnetic noise intensity, precipitation intensity, aerosol concentration, wind speed, and ambient temperature and humidity. Environmental parameters can be obtained through dedicated sensors integrated into the system, multimodal detection data inversion, or external information sources. For example, atmospheric visibility is obtained through meteorological sensors or inversion based on lidar echo attenuation rate; light intensity is determined by background brightness statistics from an ambient light sensor or infrared imaging module; aerosol concentration can be obtained from air quality data input from an environmental monitoring station; and wind speed and ambient temperature and humidity are directly collected by the anemometer, temperature sensor, and humidity sensor in the built-in meteorological micro-station. Environmental parameters can be used individually or in combination for analysis to dynamically evaluate the reliability of each module in the current environment.

[0023] This application provides an imaging device resistant to severe interference. The device is applied to the imaging system resistant to severe interference provided in any embodiment of this application. The device includes an infrared imaging module, a lidar module, and a single-photon radar module, which work together to achieve multi-dimensional data acquisition and analysis of the target environment.

[0024] The infrared imaging module is used to acquire first detection data of the target environment. This first detection data includes information on the thermal radiation distribution of the target area, which can also be referred to as infrared imaging data or infrared radiation image. An infrared imaging module is a device capable of detecting infrared radiation emitted by an object and converting it into a visual image. For example, an infrared imaging module may include an infrared detector, an optical system, and a signal processing unit: the infrared detector receives the infrared radiation emitted by the target object and converts it into an electrical signal; the optical system, composed of an infrared lens and other optical components, collects and focuses the infrared radiation emitted by the object, and its design can cover the near-infrared, mid-infrared, or long-wave infrared bands; the signal processing unit amplifies, filters, and digitizes the electrical signal output by the detector, and generates an infrared radiation image.

[0025] The lidar module is used to collect second detection data of the target environment. This second detection data includes three-dimensional spatial information and distance data of the target area, which can also be referred to as three-dimensional point cloud data. For example, the lidar module includes a laser emitting unit, a receiving unit, and a data processing unit. The laser emitting unit emits pulsed laser light of a specific wavelength, which covers the target area through a scanning device. The receiving unit includes a photodetector and a time-to-digital converter, used to receive reflected laser light and calculate the time of flight to determine the target distance. The data processing unit is used to further generate three-dimensional point cloud data.

[0026] The single-photon radar module is used to collect third-level detection data of the target environment. This third-level detection data includes high-sensitivity photon-level detection information of the target area, which is particularly suitable for low-light or weak-reflection scenarios. For example, the single-photon radar module includes a single-photon emitting unit, a single-photon detection unit, and a data processing unit: the single-photon emitting unit emits single-photon pulse signals; the single-photon detection unit records the timestamp and intensity of the single-photon reflection signal using time-correlated single-photon counting technology; and the data processing unit generates a time-correlated single-photon counting (TCSPC) histogram.

[0027] It should be noted that the specific structures of the infrared imaging module, lidar module, and single-photon radar module can be found in relevant technologies and are not limited here.

[0028] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0029] Please see Figure 1 , Figure 1 This is a schematic diagram of an imaging system resistant to severe interference provided in an embodiment of this application. Figure 1 As shown in the figure, an anti-interference imaging system provided in this application includes an infrared imaging module, a lidar module, a single-photon radar module, and a data processing unit. The infrared imaging module, lidar module, and single-photon radar module in the anti-interference imaging device are connected to the data processing unit through a system bus.

[0030] In some embodiments, the system further includes a timing controller for synchronizing the trigger signals of the single-photon radar module, the lidar module, and the infrared imaging module to align the pulse sequences of the single-photon radar module, the lidar module, and the infrared imaging module.

[0031] Specifically, the timing controller is a hardware unit used to generate and distribute synchronization signals. It generates a unified trigger signal through the master clock and sends start commands to each module to ensure that the laser pulses emitted by the single-photon radar module, the laser pulses emitted by the lidar module, and the exposure of the infrared imaging module are highly aligned at the time of emission (i.e., the pulse sequences are aligned). This achieves spatiotemporal synchronization of various types of data, so that one frame of laser point cloud corresponds to one frame of infrared image, avoiding inter-frame aliasing, thereby achieving collaborative perception and providing a consistent foundation for subsequent multimodal data fusion.

[0032] In this application embodiment, preliminary filtering is performed at both the physical and algorithmic levels when a single module collects data, in order to filter out conventional and easily identifiable interference.

[0033] In some embodiments, the single-photon radar module includes: multiple laser sources, each laser source emitting an outgoing laser signal in a preset wavelength band; each laser source having a corresponding dichroic mirror reflecting the corresponding outgoing laser signal into a polarization beamsplitter; the polarization beamsplitter combining multiple outgoing laser signals into a laser beam combiner, and deflecting the laser beam combiner by a first preset polarization angle based on a preset waveplate before emitting it; when receiving the reflected laser signal generated after the laser beam combiner is directed toward a target, the polarization beamsplitter further deflects the reflected laser signal by a second preset polarization angle based on the preset waveplate, and then splits the reflected laser signal into beams and reflects it to a corresponding single-photon detection unit based on a preset wavelength band; the single-photon detection unit records the time of each returning photon to obtain the third detection data.

[0034] Specifically, during the transmission phase, each light source emits a laser signal in a preset wavelength band. Different wavelengths of laser light are selectively reflected by a dichroic mirror to a polarizing beam splitter (PBS). The PBS combines multiple laser beams into one and adjusts the polarization direction using a preset waveplate within the PBS, ensuring the combined laser beam is emitted to the target environment at a first preset polarization angle. During the reception phase, the reflected laser signal from the target passes through the PBS again via a preset waveplate, deflecting it to a second preset polarization angle. Based on the wavelength characteristics, the beam is split to the corresponding single-photon detection unit. The single-photon detector records the arrival time of each returning photon with picosecond-level precision. By statistically analyzing the time-of-flight histogram of numerous pulses, target distance information is extracted.

[0035] In some embodiments, the number of laser sources can be flexibly configured according to the detection accuracy requirements and system cost, for example, up to three, to balance performance and economy. Furthermore, different wavelengths of laser light exhibit varying detection capabilities under different environmental conditions. In normal environments with low interference at long distances, the visible light band is helpful in acquiring the texture details of object surfaces; while in complex or harsh environments, specific infrared bands show stronger adaptability.

[0036] Preferably, the preset wavelengths of the laser source can be set to 1550nm, 2050nm, and 3500nm, respectively. 1550nm falls in the mid-infrared band, possessing good atmospheric penetration capabilities and suitable for high-scattering environments such as rain, fog, and smoke. The 2050nm band helps distinguish the reflection characteristics of different materials, improving target material identification capabilities. The 3500nm wavelength has stronger penetration and better scattering suppression performance, suitable for stable detection under extremely harsh conditions. Through multi-band collaborative operation, the system can dynamically optimize detection performance in diverse mission scenarios.

[0037] In some embodiments, a preset waveplate is used to apply specific polarization angle control to the emitted and returned beams, so that the light signals in the output direction and the received direction are orthogonal or distinguishable in polarization state. For example, the preset waveplate can be a half-waveplate, in which case the optical axis is at 45° to the polarization direction of the incident light. If the incident angle is α, and the original polarization angle of α is 0°, the polarization angle of the output light obtained after passing through the PBS is 2 × 45° - α = 90°, thus achieving separation of the incident and reflected light.

[0038] In some embodiments, a narrowband filter is provided at the entrance pupil of the single-photon radar module, which can effectively cope with strong light interference and rain and snow scattering interference, and greatly improve the noise resistance performance.

[0039] Therefore, by using polarization beam splitting technology to effectively separate the emitted laser and the echo signal, and by combining it with the setting of a narrowband filter, crosstalk of stray light and ambient background light is suppressed at the physical level, thereby reducing noise interference and improving signal effectiveness during the acquisition stage.

[0040] In some embodiments, the incident signal receiving points of the lidar module and the infrared imaging module are provided with optical phase change modules; when the power of the incident signal exceeds the corresponding anti-interference threshold, the optical phase change modules change their physical state to reduce the transmittance of the corresponding band signal.

[0041] Specifically, the imaging device integrates an optical phase-change module at the signal receiving point of the lidar module and the infrared imaging module. When strong external light or interference causes the incident signal power to exceed the anti-interference threshold, the module responds quickly and undergoes a physical state change, reducing the light transmittance of a specific band, thereby suppressing excessively strong signals from entering the detector and causing saturation or noise surge.

[0042] Among them, the optical phase change module is an optical control component based on phase change materials, which can undergo a reversible change of state under the action of external stimuli (such as temperature and light intensity), thereby dynamically adjusting the transmittance of light signals in a specific wavelength band. Correspondingly, the anti-interference threshold is the critical value of the state transition of the selected optical phase change module.

[0043] In some embodiments, the optical phase change module is a vanadium dioxide phase change module. When the power of the incident signal exceeds the corresponding anti-interference threshold, the physical state of the vanadium dioxide phase change module changes from an insulating phase to a metallic phase.

[0044] The vanadium dioxide phase change module is in an insulating phase at room temperature. When the power density of the incident signal exceeds the anti-interference threshold, the internal temperature of the module rises rapidly, triggering a transition from the insulating phase to the metallic phase. This causes changes in parameters such as transmittance and reflectivity, thereby altering its optical properties. Furthermore, the phase transition process is reversible; when the incident signal power falls below the anti-interference threshold, the vanadium dioxide phase change module can revert to its initial insulating phase. The anti-interference threshold for vanadium dioxide is the phase transition temperature, such as 65-75℃.

[0045] In some embodiments, the vanadium dioxide phase change module has been pre-processed with a metasurface design. When the laser power is too high, the vanadium dioxide will undergo a phase change and will also reflect strong light interference, thus improving its resistance to strong light interference.

[0046] In some embodiments, the optical phase transition module may also be a material with reversible phase transition properties, such as chalcogenides or nickel oxide, which can also achieve rapid switching from a transparent state to a shielded state under external stimuli such as temperature and light intensity.

[0047] Thus, by adaptively adjusting the optical phase change module, passive optical protection is achieved without relying on external control, thus physically blocking strong interference signals such as sudden strong light and direct sunlight from interfering with the lidar module and infrared imaging module.

[0048] In some embodiments, the data processing unit is further configured to: perform main peak detection based on the third detection data to determine the effective time window corresponding to the main reflection peak in the third detection data; receive the third detection data of the single-photon radar module based on the effective time window; and / or determine the target distance range based on the effective time window; filter the second detection data according to the target distance range, and retain the second detection data within the target distance range.

[0049] Specifically, the primary peak detection is performed using prior detection data collected by the single-photon radar module. The strongest signal peak, i.e., the primary reflection peak, is identified from the TCSPC histogram of the single-photon radar. This peak corresponds to the return signal of the main object in the target environment, and the time coordinate of the primary reflection peak represents the photon's flight time. The effective time window is determined based on the range of the primary reflection peak's time coordinate.

[0050] For single-photon radar modules, subsequent third-stage detection data can be received within this effective time window. It should be understood that by dynamically constructing the effective time window based on the main peak detection, only the third-stage detection data within the time interval highly correlated with the main object is retained, effectively filtering out interference from backscattered photons, multipath reflections, and random noise photons in the environmental background.

[0051] For lidar modules, the target distance range between the main object and the imaging device can be obtained by converting the flight time corresponding to the effective time window. Based on this distance range, the second detection data is filtered, retaining only the point cloud data within this range and filtering out noise points that are irrelevant to distance, thereby improving the effectiveness of the point cloud data.

[0052] In some embodiments, during the initial detection phase when the system is started or the target environment is unknown, a wider time window can be set to ensure that the complete echo signal is captured, thereby comprehensively acquiring the third detection data and accurately locating the main peak.

[0053] In some embodiments, a preset extension length is introduced as a reference parameter for extending the time window based on the main reflection peak position. Its initial value can be flexibly set according to the complexity of the actual detection environment and the stringency of data filtering. An initial effective time window is formed by superimposing the preset extension length onto the coordinate range of the main reflection peak. Based on this, an iterative optimization mode is entered, gradually reducing the preset extension length in subsequent detection cycles, so that the effective time window is gradually confined to the main peak region. This progressively eliminates backscattered photons and noise interference from other non-target time periods, and avoids misfiltering effective data.

[0054] In some embodiments, the first round initializes the time window based on the initial ranging results of the classical lidar. The single-photon radar performs peak detection on the TCSPC histogram within the time window, identifies the time point with the highest photon count as the main peak, and then adjusts the window based on the main peak's position, aligning the window's start time with the main peak to ensure the time window encompasses it. Depending on requirements, the main peak is refitted based on the signal photon count distribution within the window, or the window width is further reduced, completing subsequent iterations.

[0055] In some embodiments, in the spatial dimension, the receiving field of view can be limited by controlling the pointing angle of the scanning lens, so that only photon signals from the target direction are collected, thereby further suppressing stray light and background noise from other directions.

[0056] Therefore, at the algorithm level, the effectiveness of the detection data collected by the lidar module and the single-photon radar module is improved through inter-module collaboration, and the precise filtering of interference signals is achieved.

[0057] Please see Figure 2 , Figure 2 This is a structural schematic diagram of another imaging system resistant to severe interference provided in an embodiment of this application. For example... Figure 2As shown, in the anti-interference imaging system 100, the infrared imaging module can be a long-wave infrared imaging module 10, mainly composed of an infrared camera 11, a variable-focus infrared lens 12, and a phase-change module 40 resistant to strong light interference. The variable-focus infrared lens 12 is the optical system of the infrared imaging module; by adjusting the focal length, it clearly captures the infrared information of objects at different distances, improving detection flexibility and imaging quality. The lidar module can be a classic lidar module 20, including a classic lidar and a phase-change module 40 resistant to strong light interference. The single-photon radar module is a single-photon-level ultrafast lidar module 30, mainly composed of: three small picosecond laser sources of different wavelengths LD1, LD2, and LD3; three coated dichroic mirrors DM1, DM2, and DM3 corresponding to the wavelengths on three sides; a polarization beam splitter PBS; a two-dimensional scanning prism GS; a broadband high-reflection mirror BM; three narrowband filters OF corresponding to the wavelengths; and three time-dependent counter-coupled time-shutter single-photon avalanche diode (SPAD) cameras TSC.

[0058] During the detection process, the timing controller 50 sends trigger signals to the long-wave infrared imaging module 10, the classical lidar module 20, and the single-photon-level ultrafast lidar module 30 respectively to align the pulse sequence and unify the t0 timestamp.

[0059] In the long-wave infrared imaging module 10, the variable-focus infrared lens 12 receives the thermal signal radiated by the target. The timing controller 50 triggers the detector FPA (Focal Plane Array) to start exposure. The exposure time is synchronized with the pulse period of the classic lidar 20. The long-wave infrared imaging module 10 collects the thermal radiation signal after passing through the phase change module 40 which is resistant to strong light interference, completes the exposure, and generates an infrared image with a timestamp t0 (i.e., the first detection data), which is synchronously transmitted to the data processing unit 60 (such as an edge computing unit).

[0060] In the classic lidar module 20, a pulsed laser source emits a laser beam, which is scanned horizontally and vertically by a scanning lens to form a fan-shaped detection beam directed towards the target scene. The flight time of the laser pulse is marked as time t0 by the timing controller 50. The echo signal of the classic lidar first passes through the phase transition module 40, which is designed to resist strong light interference. When the incident laser power increases to the anti-interference threshold, it rapidly transforms into a metallic state within an ultra-short time of less than picoseconds, causing the transmittance of the corresponding wavelength band to drop rapidly by nearly 10 orders of magnitude. This prevents the sensor from oversaturating, thereby achieving the purpose of anti-interference and protection. The remaining signal is focused by the receiving lens, converted into an electrical signal by the APD detector, and the laser flight time Δt is measured by a time-to-digital converter. Combined with the scanning galvanometer angle, the three-dimensional coordinates of the target are calculated to generate point cloud data (i.e., the second detection data) and marked with a timestamp of t0+Δt.

[0061] In the single-photon ultrafast lidar module 30, three picosecond laser sources, LD1, LD2, and LD3, simultaneously emit lasers. The lasers are reflected by dichroic mirrors DM1, DM2, and DM3, respectively, and then input into the polarization beam splitter PBS for beam combining. The combined single-photon laser beam passes through a half-wave plate and is directed towards the target environment. The reflected light from the target environment passes again through the half-wave plate, undergoes a 90° polarization rotation, and is reflected by the polarization beam splitter PBS to the single-photon detector SPAD array. Wavelength beam splitters separate the echoes into different wavelength bands, recording the timestamps and spatial locations of photon events to obtain the third detection data.

[0062] Through the coordinated operation of the aforementioned modules, this system combines the thermal radiation characteristics of the infrared imaging module, the three-dimensional spatial perception capability of the lidar module, and the high-sensitivity detection of the single-photon radar module to achieve multi-dimensional, all-weather target monitoring in complex environments. It should be understood that in high-scattering environments such as rain, fog, and dust storms, the detection range and accuracy of lidar and single-photon radar decrease, but infrared imaging, based on thermal radiation, is less affected and can provide stable target outlines. When facing light pollution such as sunlight and searchlights, specific band signals of lidar and single-photon radar are easily submerged by noise, but the mid-to-long-wave bands of infrared imaging are relatively less affected by interference and can provide background information. When encountering thermal radiation interference such as thermal decoys, infrared images may be deceived, but the precise three-dimensional point cloud data provided by lidar and single-photon radar, due to their completely different detection mechanisms, are unaffected by thermal radiation and can effectively distinguish between real and false targets. This multi-source capability allows the system to obtain crucial information through other sensors even if one sensor fails, fundamentally improving robustness.

[0063] Furthermore, the first detection data collected by the infrared imaging module, the second detection data collected by the lidar module, and the third detection data collected by the single-photon radar module are transmitted to the data processing unit 60 to perform dynamic weighted filtering and imaging in the processing stage.

[0064] Please see Figure 3 , Figure 3 This is a flowchart illustrating the data processing process of a data processing unit according to an embodiment of this application, such as... Figure 3 As shown, in this embodiment of the application, the data processing unit in the imaging system is used to perform steps S201 to S204.

[0065] S201, acquire the first detection data collected by the infrared imaging module, the second detection data collected by the lidar module, and the third detection data collected by the single-photon radar module.

[0066] S202, adjust the weights among the first detection data, the second detection data, and the third detection data according to the environmental parameters of the target environment; wherein, the weights include the confidence weight of the first detection data and the fusion weight of the second and third detection data.

[0067] Specifically, the degree of interference experienced by the infrared imaging module, single-photon radar module, and lidar module varies under different harsh environments. The weights of the first, second, and third detection data are dynamically adjusted according to the environmental parameters of the target environment to optimize the fusion imaging effect.

[0068] In some embodiments, the reliability of the first detection data output by the infrared imaging module is evaluated based on the target environmental parameters, and a corresponding confidence weight is assigned to it. For example, in strong light or electromagnetic interference environments, if the infrared image is clear and stable, the confidence weight is maintained at a high level, while in the presence of thermal interference such as thermal decoys, its weight is reduced.

[0069] In some embodiments, the confidence weight of the first probe data is not set in isolation, but is configured collaboratively based on its relative reliability relationship with the second and third probe data. Based on this weight allocation, the preset error thresholds corresponding to the second and third probe data can be adaptively fine-tuned according to their respective weight proportions and set independently. For example, when the confidence weight of the third probe data is higher than the standard weight value, its preset error threshold is correspondingly lower, and the filtering strategy is relatively lenient; simultaneously, when the confidence weight of the second probe data is lower than the standard weight value, its preset error threshold is higher, and the filtering strategy is relatively strict.

[0070] In some embodiments, an association mapping table of environmental parameters and weights is pre-set based on prior data, and the initial standard weight values ​​are adaptively optimized for different target environment environmental parameters.

[0071] For example, the standard weight values ​​can be configured as: "First probe data: Second probe data: Third probe data = 0.1:0.5:0.4".

[0072] For example, if atmospheric visibility is below a preset threshold, it indicates the presence of dense fog, haze, or dust storms, and the rain sensor displays moderate to heavy rain, suggesting severe weather interference. In this situation, enhanced scattering of short-wavelength light signals leads to increased distance deviation and decreased confidence in the single-photon radar module. While the continuous point cloud information from the lidar module and the infrared imaging data from the infrared imaging module remain highly reliable, the infrared imaging data, used as an auxiliary criterion, should not have excessively high weight. Therefore, the confidence weight can be configured as: "First detection data: Second detection data: Third detection data = 0.1:0.55:0.35".

[0073] For example, if the light intensity is significantly higher than normal, environmental parameters indicate light pollution. Strong external light sources (such as direct sunlight or city searchlights) partially overlap with single-photon and laser wavelengths, causing a surge in noise in single-photon detectors. Weak signal points in the lidar are easily overwhelmed, reducing the confidence levels of both the single-photon radar module and the lidar module. Consequently, their weights are adjusted downwards. In this case, the confidence weights can be configured as: "First detection data: Second detection data: Third detection data = 0.2:0.45:0.35".

[0074] For example, if there is electromagnetic noise in the environment that exceeds the normal background level, and this noise has a high field strength in a specific frequency band (such as radio frequency, intermediate frequency, or power frequency), the environmental parameters indicate an electromagnetic interference environment. Electromagnetic radiation may cause counting errors in the time-to-digital converter of the single-photon radar module or phase / pulse detection errors in the laser ranging module, leading to a decrease in the data reliability of the single-photon radar module. In this case, the confidence weight can be configured as: "First detection data: Second detection data: Third detection data = 0.1: 0.55: 0.35".

[0075] For example, if a localized high-heat area with abnormal thermal radiation levels appears in the environment, the environmental parameters indicate a thermal radiation interference environment. If a target area contains strong infrared interference sources such as thermal decoys, causing distortion or even complete failure of the infrared image contour, and if a region has significant thermal radiation features in its infrared contour, but the corresponding location does not have a high-confidence point cloud response in single-photon radar or lidar data, it is determined to be thermal decoy interference. In this case, the weight of the first detection data is reduced to the minimum and can be configured as: "First detection data: Second detection data: Third detection data = 0.05: 0.5: 0.45".

[0076] In some embodiments, the fusion weights of the second and third detection data within different detection ranges are determined by comprehensively considering the detection performance and ranging characteristics of lidar and single-photon radar in different environments. The fusion weights are dynamically set coefficients based on the detection range and environmental parameters, used to determine the relative contribution ratios of the second and third target detection data during the fusion process.

[0077] For example, fusion weights can be set for different detection ranges and environments, such as short-range, medium-range, and long-range: in the medium-range high-scattering environment, the data of the lidar module is emphasized, and the fusion weight can be configured as: "Second detection data: Third detection data = 0.7: 0.3"; in the long-range clear conditions, the data weight of the single-photon radar module is enhanced, and the fusion weight can be configured as: "Second detection data: Third detection data = 0.5: 0.5".

[0078] For example, in an electromagnetic interference environment, the data reliability of a single-photon radar module decreases. In this case, the confidence weight can be configured as: "Second detection data: Third detection data = 0.6: 0.4".

[0079] It should be understood that the embodiments of this application construct a dynamic filtering link from direct filtering to difference selection through a hierarchical design of two types of weights, so as to eliminate interference data and retain effective information, while taking into account the complementarity of different sensors in terms of environmental adaptability.

[0080] The first type of weight is used for direct filtering based on confidence level. The confidence level weight of the infrared imaging module is directly used as the basis for setting the filtering threshold. The reliability of the first detection data is dynamically evaluated through environmental parameters, and other detection data are screened based on this. This achieves adaptive adjustment of the error threshold, enabling rapid elimination of interference data while retaining detection data with strong environmental adaptability. For example, in a heat decoy interference scenario, the confidence level weight of the infrared data is lowered, causing the generated contour features to relax the verification standard of lidar and single-photon radar data (i.e., the preset error threshold is increased), thereby avoiding over-filtering of real target information due to false heat source misjudgment. In a high scattering environment, the confidence level weight of the infrared data is higher, and its contour features become a strict verification standard for lidar point clouds and single-photon radar echo data, directly filtering out abnormal points that deviate from the contour.

[0081] The second type of weight is used for difference selection based on the fusion strategy. The fusion weights of the second and third detection data are allocated differentially according to environmental and spatial dimensions to selectively use the differences between the two types of data. Essentially, this is also part of the filtering mechanism. By dynamically adjusting the contribution ratio of the two types of detection data, the influence of low-confidence data caused by environmental characteristics (such as scattering intensity and noise level) is suppressed. For example, in a close-range, high-scattering environment, single-photon radar has a higher fusion weight due to its high sensitivity and resistance to weak signals, and it dominates the generation of the target outline.

[0082] The first type of weight uses confidence assessment driven by environmental parameters to remove interference data at the global level, while the second type of weight adapts to differences in data quality within the local detection range through dynamic fusion of environmental and spatial dimensions. By decoupling the confidence weight and the fusion weight, corresponding to the two dimensions of environmental interference filtering and sensor performance optimization respectively, a more refined filtering strategy is formed. This avoids over-filtering due to sudden environmental changes (such as using a lidar with fixed weights after attenuation in rainy weather) or misjudgment (such as not dynamically adjusting the verification standard of infrared data under thermal decoy interference) caused by using fixed weights or a single confidence threshold.

[0083] S203, a first contour feature is generated based on the first detection data, and the second detection data and the third detection data are filtered according to the first contour feature and a preset error threshold to obtain target second detection data and target third detection data; wherein, the preset error threshold is determined according to the confidence weight.

[0084] Specifically, the thermal radiation contour of the target is extracted based on the first detection data output by the infrared imaging module, generating a first contour feature to characterize the approximate boundary and shape distribution of the target object in space. The second and third detection data acquired by the lidar and single-photon radar are spatially aligned with the first contour feature, and a preset error threshold is used to determine whether the detection data is within a reasonable deviation range. Data exceeding the error threshold is identified as background noise or interference signals and is removed, thereby obtaining cleaner second and third target detection data.

[0085] The preset error threshold is not a fixed value, but is dynamically adjusted based on the confidence weight of the first detection data. For example, in a nighttime environment with good visibility, the infrared imaging is clear and the confidence weight is high, so the error threshold is set to a small value (e.g., ±0.1 meters) to achieve high-precision filtering. However, when there is interference from thermal decoys or dense fog, the reliability of the infrared data decreases and the confidence weight decreases, so the error threshold is increased accordingly (e.g., ±0.5 meters) to avoid misfiltering real targets due to contour distortion, thereby improving the system's fault tolerance and robustness in complex environments.

[0086] The first contour feature is the target thermal radiation boundary information extracted from infrared imaging data.

[0087] S204, based on the fusion weight, the second target detection data and the third target detection data are fused to generate a target image.

[0088] Specifically, based on pre-determined fusion weights, the target's second and third detection data, after contour filtering, are weighted and fused to generate a high-precision, robust target image. The fusion process is not a simple superposition; rather, based on detection performance and ranging characteristics, the two types of data are aligned in three-dimensional space and their contribution proportions are allocated according to their weights. For example, in a rainy urban night scene, the contours of nearby vehicles are primarily completed using the second detection data from the LiDAR module, avoiding raindrop noise interference, while the road structures at medium and long distances are filled in using the third detection data from the single-photon radar module, ensuring the overall image's coherence and accuracy. Through a dynamically weighted fusion strategy, the system adaptively optimizes imaging quality in different regions and environments, improving overall perception reliability.

[0089] Thus, the obtained target image is a three-dimensional spatial representation image generated from the fused multi-source detection data, containing the geometric structure, distance information and spatial distribution of the target object, which is used for subsequent identification and decision-making.

[0090] It should be understood that dynamic weight filtering occurs during the processing phase. The system dynamically adjusts the confidence weight and fusion weight of the detection data based on environmental parameters to adapt to the different modules' adaptability in various harsh environments. Then, using the infrared data to generate the contour as a benchmark, a dynamic error threshold corresponding to the confidence weight is set to verify and filter the laser and single-photon point clouds.

[0091] Furthermore, a reverse verification and compensation mechanism is introduced: for abnormally missing continuous contours, the error threshold is temporarily relaxed, and the suspicious data is re-verified and the reasonable contours are completed. The closed-loop strategy of "filtering-verification-compensation" retains key target information while avoiding the introduction of interfering data, ultimately achieving reliable target reconstruction in complex interference environments.

[0092] In some embodiments, please refer to Figure 4 , Figure 4 This is a flowchart illustrating the data processing process of another data processing unit provided in this application embodiment, such as... Figure 4 As shown, in this embodiment of the application, the data processing unit in the imaging system is also used to perform steps S301 to S303.

[0093] S301, based on the second target detection data and the third target detection data, generate corresponding second contour features and third contour features respectively.

[0094] The second contour feature is a geometric representation generated based on the second detection data of the target through methods such as boundary extraction, geometric fitting, or morphological processing. The third contour feature is a geometric representation generated based on the third detection data of the target through methods such as signal strength analysis, temporal correlation, or spatial clustering. It should be understood that the first, second, and third contour features are identifiable shape structures generated under different detection mechanisms; specific generation methods can be found in relevant technologies.

[0095] S302, if a local missing continuous contour is detected in a local area of ​​the second contour feature and / or the third contour feature, the local area to be verified detection data is updated based on the temporary error threshold, and the contour feature of the local area is updated based on the local area to be verified detection data.

[0096] In this context, a continuous contour is composed of multiple ordered connected feature points or line segments. A certain segment in the continuous contour may have its feature information interrupted or data blanked due to occlusion, signal attenuation or interference, or excessive filtering, resulting in a local break or sparse feature in the continuous contour, i.e., a local missing feature in the continuous contour.

[0097] In some embodiments, if a local missing section of a continuous contour is detected in the same local region of the second contour feature and the third contour feature, the preset error threshold is updated to a temporary error threshold. Here, the same local region refers to a region that coincides or highly overlaps in the spatial coordinate system, used to align and compare the observation results of different sensors (such as lidar and single-photon radar) at the same target location.

[0098] The temporary error threshold is a dynamically activated temporary tolerance threshold with a value greater than the preset error threshold. It is used to relax the judgment conditions for the deviation between the second and third detection data of the target and the reference first contour feature in areas with missing local data, allowing more potential effective detection points to be included in the analysis scope and avoiding false filtering.

[0099] S303, if the contour features of a local area meet a preset continuity standard, the detection data to be verified is updated to target second detection data and / or target third detection data.

[0100] Among them, the preset continuity standard is a criterion for judging whether the contour meets the requirements of integrity and smoothness, including but not limited to: the contour line is unbroken within a certain length range, the curvature change is continuous and does not exceed the threshold, the directional angle between adjacent segments is gentle, and the interpolation can form a closed or connected path, etc.

[0101] Specifically, based on the second and third target detection data, corresponding second and third contour features are generated respectively. By comparing parameters such as distance and angle changes between contour segments, it is determined whether there is a local missing section of the continuous contour. If so, the second and / or third detection data in that area are re-filtered according to the first contour feature and a temporary error threshold to obtain the detection data to be verified. Based on this, the contour features in that area are regenerated, and it is determined whether the updated contour features meet a preset continuity standard, i.e., the contour remains smooth and without abrupt changes within a specific length range. If it meets the standard, it is determined that there has been false filtering, and the detection data to be verified is updated with the target's second and / or third detection data for that area, thereby improving the overall detection accuracy and completeness. If it does not meet the standard, it is determined that feature information is missing at that location due to occlusion, signal attenuation, interference, or other reasons, and the detection data to be verified is filtered out.

[0102] In some embodiments, the data processing unit is further configured to: perform skeletonization processing on the second target detection data and the third target detection data to obtain the second contour feature and the third contour feature, wherein the second contour feature is composed of a plurality of second skeleton points and the third contour feature is composed of a plurality of third skeleton points; when the second skeleton point and / or the third skeleton point in the same region is detected to be missing, and the skeleton points in adjacent regions are continuous, confirm that a local missing part of the continuous contour has occurred; determine the skeleton point to be verified based on the detection data to be verified, and update the contour feature based on the skeleton point to be verified and the skeleton points in adjacent regions.

[0103] Specifically, skeletonization processing is performed on the filtered target second and third detection data. For details, please refer to relevant technologies. Related algorithms generate second and third skeleton points, which are sequentially connected to form second and third contour features. When both second and third skeleton points are simultaneously missing in the same spatial region, or when either the second or third skeleton point is missing in any region, but the skeleton points in adjacent regions remain continuously distributed, it is determined to be a local missing section of a continuous contour. For example, in vehicle detection, if no valid data is detected in the middle of the vehicle after filtering, but the skeleton point chains at the front and rear of the vehicle are complete.

[0104] The skeleton points corresponding to the detection data to be verified are used as the skeleton points to be verified, and contour features are regenerated with the skeleton points in the adjacent regions. Further, step S304 is executed to verify the continuity of the contour features.

[0105] In some embodiments, the preset continuity criteria may include the uniformity of spacing between skeleton points, the continuity of directional changes, and the smoothness of curvature changes. Local defects refer to the phenomenon of a large number of skeleton points being missing or the signal being interrupted in a certain segment of a continuous contour.

[0106] In some embodiments, the temporary error threshold is greater than the preset error threshold; the data processing unit is further configured to: update the temporary error threshold to the preset error threshold to achieve reverse correction.

[0107] The temporary error threshold is set based on a comprehensive evaluation of current environmental parameters and the confidence weights of multiple sensors. It is suitable for short-term, localized abnormal interference scenarios, such as occlusion, strong scattering, or signal attenuation areas. Once the contour is repaired and meets the preset continuity standard, the system updates the temporary error threshold back to the original preset error threshold, restores the regular filtering strategy, and ensures a balance between overall detection accuracy and robustness.

[0108] It should be understood that the preset error threshold is dynamically determined by the confidence weight of the first detection data. It is a global and normalized filtering standard, reflecting the system's tolerance requirements under current harsh environmental conditions. The temporary error threshold is generated based on the preset error threshold. Its value is obtained by adding a preset value to the current preset error threshold. It is a localized and temporary moderately relaxed filtering standard for a specific area, reducing the data filtering intensity, identifying valid detection points that may be misfiltered under the preset error threshold, and then using the detection data to be verified for contour completion.

[0109] The temporary error threshold is generated based on the preset error threshold, inheriting its environmental adaptive characteristics and expanding the tolerance range on this basis. It maintains the controllability of overall data quality through the preset error threshold, and realizes the fault tolerance perception of specific areas through the temporary error threshold, avoiding information loss caused by a single strict standard, and improving the integrity and adaptability of the system in complex and dynamic environments.

[0110] In some embodiments, the preset error threshold value can be dynamically adjusted within a preset upper and lower limit range according to the change of confidence weight, so as to avoid the filtering strategy being too strict (such as mistakenly removing true points) or too lenient (such as introducing too much noise) due to the change of weight, thereby ensuring the stability and reliability of the system in various environments.

[0111] In some embodiments, echo signals from laser generators and infrared lenses of different wavelengths are comprehensively utilized to collect multimodal data, including point cloud data from classical lidar, wavelength data from infrared imaging, and TCSPC data from single-photon radar. Data augmentation is performed through computational modeling using neural networks and thorough analysis and cross-validation of the multimodal data, fully exploring the correlation between echo signals of different wavelengths and significantly improving radar detection capabilities. This system functions effectively under complex conditions such as artificial smoke, severe natural weather, and photoelectric electromagnetic interference. By analyzing echo signals containing time series data, accurate 3D imaging can be achieved, enabling the detection, tracking, and perception of different types of dynamic targets.

[0112] In some embodiments, after processing by a confocal beam combiner and polarization beam splitter system, the echo signal of the single-photon laser is separated into corresponding detection channels for each band by a narrowband filter. The echo signals of classical laser and infrared radiation pass through an anti-strong light interference phase transition module and are then connected to a high-speed acquisition system, simultaneously capturing weak single-photon level signals, point cloud data signals, and thermal radiation signals. Each module achieves spatiotemporal alignment through a unified timestamp. Using a deep learning model from the edge computing module, the multimodal heterogeneous data collected by different modules are fully fused to reconstruct the target image within a limited field of view, thus realizing a typical dynamic target detection and tracking task under complex environmental interference.

[0113] In some embodiments, single-photon-level sensitivity and strong penetration enable the capture of weak signals and suppression of environmental noise, significantly improving the system's detection range and anti-interference capabilities. Simultaneously, a target perception enhancement model for complex scenes is constructed, achieving cross-modal feature coupling between the photon time distribution histogram obtained from single-photon radar and heterogeneous features such as classical radar point clouds and infrared thermal radiation information. This fully integrates heterogeneous data to reconstruct three-dimensional images of objects, greatly enhancing data readability and utilization.

[0114] In some embodiments, the fusion principle of the 3D object imaging algorithm is as follows: Relevant features are extracted from various types of detection data through a three-stream CNN network. Object reconstruction is mainly achieved using the second and third target detection data, with the first detection data serving as auxiliary semantic features in the fusion process. Single-photon radar can still effectively acquire depth data in scenarios such as low light, long-range detection, or low target reflectivity, compensating for the performance degradation of classical lidar under these conditions. Classic lidar, on the other hand, can provide high-density point clouds with accurate geometric features and strong real-time performance, thus compensating for the data sparsity and weak real-time performance of single-photon radar. Infrared telephoto lenses possess thermal feature recognition capabilities, support all-weather operation, and can assist in identifying camouflaged or partially obscured targets, solving the recognition difficulties caused by similar shapes or visual confusion in classical radar. Infrared imaging lacks accurate depth information and has a low recognition rate for objects with similar temperatures; this deficiency can be compensated for by the geometric structural information of lidar and single-photon radar. Through fusion, the long-range detection capability of single-photon radar and the short-range point cloud density of classical lidar together achieve full-range coverage; the fine geometric features provided by radar and the thermal profile features generated by infrared imaging together constitute a full-granular characterization of the target. At the same time, single-photon radar has the ability to resist weak light interference, infrared imaging has the ability to resist strong light interference, and classical lidar has high real-time performance. The three work together to achieve all-weather, robust perception capabilities.

[0115] The anti-interference imaging system provided in this application constructs a hierarchical anti-interference system from data source acquisition to back-end fusion processing. Through a collaborative verification mechanism, data from different modalities can complement and verify each other when facing specific interference, thereby outputting clear and coherent reconstructed images of the object under test in extremely complex integrated environments, improving the perception capability and anti-interference performance of real targets.

[0116] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An imaging system resistant to severe interference, characterized in that, The system includes an infrared imaging module, a lidar module, a single-photon radar module, and a data processing unit; The data processing unit is used for: Acquire the first detection data collected by the infrared imaging module, the second detection data collected by the lidar module, and the third detection data collected by the single-photon radar module; Based on the environmental parameters of the target environment, the weights among the first detection data, the second detection data, and the third detection data are adjusted; wherein, the weights include the confidence weight of the first detection data and the fusion weight of the second and third detection data; A first contour feature is generated based on the first detection data, and the second and third detection data are filtered according to the first contour feature and a preset error threshold to obtain target second detection data and target third detection data; wherein, the preset error threshold is determined according to the confidence weight; Based on the fusion weights, the second and third target detection data are fused to generate a target image.

2. The system as described in claim 1, characterized in that, The data processing unit is also used for: Based on the second and third target detection data, corresponding second and third contour features are generated respectively; If a local missing continuous contour is detected in a local area of ​​the second contour feature and / or the third contour feature, the detection data to be verified in the local area is updated based on the temporary error threshold, and the contour features of the local area are updated based on the detection data to be verified. If the contour features of a local area meet a preset continuity standard, the detection data to be verified will be updated to target second detection data and / or target third detection data.

3. The system as described in claim 2, characterized in that, The data processing unit is also used for: The second and third target detection data are processed into skeletonization to obtain the second contour feature and the third contour feature. The second contour feature is composed of multiple second skeleton points and the third contour feature is composed of multiple third skeleton points. When the second and / or third skeleton points in the same region are detected to be missing, and the skeleton points in adjacent regions are continuous, it is confirmed that there is a local missing part of the continuous contour. The skeleton points to be verified are determined based on the detection data to be verified, and the contour features are updated based on the skeleton points to be verified and the skeleton points in the adjacent regions.

4. The system as described in claim 2 or 3, characterized in that, The temporary error threshold is greater than the preset error threshold; The data processing unit is also used to update the temporary error threshold to a preset error threshold.

5. The system as described in claim 1, characterized in that, The data processing unit is also used for: The main peak is detected based on the third detection data to determine the effective time window corresponding to the main reflection peak in the third detection data. Based on the effective time window, the third detection data of the single-photon radar module is received; And / or, The target distance range is determined based on the effective time window; the second detection data is filtered according to the target distance range, and the second detection data within the target distance range is retained.

6. The system as described in claim 1, characterized in that, The single-photon radar module includes: Multiple laser sources, each of which is used to emit an outgoing laser signal in a preset wavelength band; Each of the laser sources is provided with a corresponding dichroic mirror, which is used to reflect the corresponding emitted laser signal to the polarization beam splitter; The polarization beam splitter is used to combine multiple outgoing laser signals into a laser beam combiner signal, and to deflect the laser beam combiner signal by a first preset polarization angle based on a preset waveplate before emitting it. When receiving the reflected laser signal generated after the laser beam combining signal is directed toward the target, the polarization beam splitter is also used to deflect the reflected laser signal by a second preset polarization angle based on the preset waveplate, and then split the reflected laser signal and reflect it to the corresponding single-photon detection unit based on the preset band. The single-photon detection unit is used to record the time of each returning photon to obtain the third detection data.

7. The system as described in claim 1, characterized in that, The incident signal receiving points of the lidar module and the infrared imaging module are equipped with optical phase change modules. When the power of the incident signal exceeds the corresponding anti-interference threshold, the optical phase change module changes its physical state to reduce the transmittance of the signal in the corresponding band.

8. The system as described in claim 7, characterized in that, The optical phase change module is a vanadium dioxide phase change module. When the power of the incident signal exceeds the corresponding anti-interference threshold, the physical state of the vanadium dioxide phase change module changes from an insulating phase to a metallic phase.

9. The system as described in claim 1, characterized in that, The system also includes a timing controller, which is used to synchronize the trigger signals of the single-photon radar module, the lidar module, and the infrared imaging module, so that the single-photon radar module, the lidar module, and the infrared imaging module align the pulse sequences.

10. An imaging device resistant to severe interference, characterized in that, The device is used in an imaging system resistant to severe interference as described in any one of claims 1 to 9, the device comprising: The infrared imaging module is used to collect the first detection data corresponding to the target environment; The lidar module is used to collect secondary detection data corresponding to the target environment; The single-photon radar module is used to collect third-party detection data corresponding to the target environment.

Images