Unmanned cooperative detection method and system based on SAR guided optical load

Abstract

The invention discloses an unmanned cooperative detection method and system based on an SAR guided optical load. The method comprises the following steps: S1, acquiring an echo signal through an SAR antenna and demodulating the echo signal; s2, inputting the demodulated echo signal into an FPGA (Field Programmable Gate Array) to obtain an SAR image; s3, transmitting the SAR image to an edge computing platform; s4, target detection is carried out on the basis of YOLO-TensorRT; judging whether a target is detected or not; s5, determining a final target through confidence coefficient screening and circumcircle geometric verification; s6, performing conversion from the pixel coordinates to the geographic coordinates to obtain the WGS-84 geographic coordinates of the final target; s7, the unmanned aerial vehicle receives an approaching flight path instruction of the ground station and flies to the target area; and the optical load starts a high-definition imaging mode to verify target details. According to the method, the technical problems of poor real-time performance, low resource cooperation efficiency, inaccurate target guidance and the like in multi-sensor cooperative detection in a complex scene can be solved.

Description

Technical Field

[0001] The present invention relates to the fields of unmanned aerial vehicle detection and synthetic aperture radar image processing, and particularly to an unmanned collaborative detection method and system based on SAR-guided optical payloads. Background Art

[0002] The leapfrog development of unmanned technology has promoted its in-depth application in strategic fields such as military reconnaissance, disaster monitoring, and border patrol. However, the target detection task in complex scenarios is facing multi-dimensional technical challenges. From the perspective of environmental adaptability, the electromagnetic multipath effect of urban building complexes, the shielding interference of dense forest vegetation, and the elevation change of mountainous terrain make it difficult for a single sensor to meet the actual detection needs. In this technical context, multi-sensor collaborative detection has become the core path to break through the performance bottleneck. Among them, the collaboration between synthetic aperture radar (SAR) and optical payloads shows unique advantages: SAR is an active microwave remote sensing technology that can obtain all-weather, all-day, and large-area SAR images. In recent years, the development of miniature SAR technology can support real-time imaging in a wide area. On the other hand, optical payload pods represented by visible light / infrared can provide high-resolution imaging and can accurately identify targets, but their viewing angle range is narrow and the resolution is insufficient. The collaborative mode based on SAR-guided optical payloads can theoretically balance the wide-area search efficiency and target recognition accuracy, but there are still key technical bottlenecks such as hardware collaboration and real-time processing in the actual engineering process.

[0003] First, there are problems of real-time performance and resource collaboration efficiency in the collaborative detection between SAR and optical payloads. SAR imaging needs to process a large amount of raw data. If it is transmitted back, it will compete with flight control commands within the limited bandwidth. Traditional imaging, positioning, and recognition algorithms are all deployed on the same hardware platform, which is prone to resource overload and performance fluctuations at high temperatures. Therefore, how to design a heterogeneous computing architecture and dynamic resource scheduling to achieve efficient collaboration has become the key to breaking through the real-time bottleneck and ensuring the stable operation of the system. Looking back at the existing related patent technologies, CN119289984A proposed a multi-payload task planning optimization algorithm, but its core limitation is that it did not build an efficient hardware collaboration architecture and only coordinated the work of sensors through traditional path planning, unable to solve the problems of real-time processing and resource competition. CN119339067A focuses on SAR-optoelectronic collaboration, but it did not design a dedicated hardware collaboration architecture and relies on SAR post-processed images. The current patent technologies related to the data fusion of SAR and optical payloads (CN119762557A proposed visible light and SAR registration, CN119762558A proposed infrared and SAR registration) have built a registration and fusion algorithm framework based on multi-scale descriptors, but they did not build a hardware collaboration architecture and have limitations in engineering adaptability.

[0004] It can be seen that although the existing patent technologies have achieved task planning, data fusion, etc. for SAR and optical payloads, they only focus on the data processing flow at the algorithm level, and do not fully consider the hardware cooperation mechanism, dynamic environment adaptation and engineering implementation requirements in the actual application of UAVs. There are problems such as insufficient real-time performance, inefficient resource scheduling, and poor robustness in complex scenarios. Summary of the Invention

[0005] In view of the above, the present invention proposes an unmanned cooperative detection method and system based on SAR-guided optical payloads. Through a full-process real-time processing chain, a closed loop from SAR raw echo acquisition to target geographical coordinate guidance is realized, and technical problems such as poor real-time performance, low resource cooperation efficiency, and inaccurate target guidance in multi-sensor cooperative detection in complex scenarios are solved.

[0006] The present invention proposes an unmanned cooperative detection method based on SAR-guided optical payloads, including the following steps: S1: The SAR antenna performs strip scanning on the target area to obtain echo signals, and the echo signals are demodulated to obtain the demodulated echo signals; S2: The demodulated echo signals are input into the FPGA for range and azimuth compression to achieve imaging, and SAR images are obtained; S3: The SAR images are transmitted to the edge computing platform through the COFDM data transmission link of the data transmission communication radio; S4: The lightweight YOLOv11 model is quantized to INT8 to obtain YOLO-TensorRT; target detection is performed on the SAR images based on YOLO-TensorRT; it is judged whether a target is detected: when there is a target, S5 is performed; if no target is detected, return to S1; S5: Confidence screening and circumscribed circle geometric verification of the detection results of target detection are performed to determine the final target; S6: Combining the SAR images, the circumcenter coordinates of the circumscribed circle of the final target are used as the input for the conversion from pixel coordinates to geographical coordinates, the conversion from pixel coordinates to geographical coordinates is performed, and elevation correction is performed through digital elevation model data to obtain the WGS-84 geographical coordinates of the final target; S7: The edge computing platform transmits the WGS-84 geographical coordinates of the final target to the ground station through the ground base station. The UAV receives the approaching flight path instructions forwarded by the ground station through the ground base station, adjusts the heading and altitude, and flies towards the target area; the optical payload starts the high-definition imaging mode to verify the target details.

[0007] Further, the S1 includes: S11: The SAR antenna transmits radar signals to the target area, and the expression of the echo signal after being reflected by the target and received is as follows: ; Among them, represents the echo signal reflected by the target received by the SAR; is the target reflection coefficient, which is a complex constant, t, respectively represent the slow time dimension and the fast time dimension, is the moment when the beam center passes through the target, and c is the propagation speed of electromagnetic waves; represents the azimuth signal reception intensity, and a represents the azimuth, that is, the direction related to the SAR movement direction; represents the signal range pulse envelope intensity; is the slant range between the SAR and the target at time t; exp() is the exponential function; is the center frequency, is the range modulation frequency; and respectively represent the signal range pulse envelope function and the azimuth signal reception intensity function, and their expressions are as follows: ; Among them, is the pulse duration, P a is the azimuth antenna pattern function, rect() is the rectangular function, represents when ; sinc is the sinc function; is the angle between the target and the beam center line at time t; is the azimuth beam width, is the signal wavelength, is the azimuth antenna width; S12: Demodulate the echo signal to obtain the demodulated echo signal: ; Among them, represents the demodulated echo signal; j is the imaginary unit.

[0008] Furthermore, the S2 includes: Use the demodulated echo signal as the SAR raw data and input it into the FPGA to perform range and azimuth compression to achieve imaging. The specific steps are as follows: S21: Range compression; Perform FFT on the range direction of the demodulated echo signal to obtain the following expression: ; Among them, is the echo signal after FFT processing in the range direction of the demodulated echo signal; G is the total gain including the scattering coefficient, set to 1; is the envelope of the range spectrum, is the range frequency; is the slant range between the SAR and the target at slow time η; Range-matching filter The expression is: ; The echo signal after FFT processing is output through the range filter and then through IFFT to obtain the expression after range compression : ; Among them, is the inverse fast Fourier transform; is The IFFT expression of; S22: Azimuth FFT; The slant range of the SAR from the target : ; Among them, is the initial slant range, v is the SAR platform speed; Since the beam points close to the zero Doppler direction, and , approximate as a parabola, the slant range of the SAR from the target , the calculation formula can be replaced by: ; When performing azimuth FFT, the time-frequency relationship is , among them, is the azimuth frequency; is the azimuth chirp rate; is the SAR platform movement speed; is the wavelength of the SAR transmitted signal; the azimuth FFT expression can be obtained as follows: ; Among them, is the echo signal after azimuth FFT processing; is the fast Fourier transform of the time variable t; is the echo signal after range compression; is the azimuth frequency variable; is The frequency domain form of; That is, the range migration compensation term, and its expression is as follows: ; S23: Perform range migration correction; Range migration amount to be corrected , the expression is as follows: ; Range migration correction factor , the expression is as follows: ; Then the echo signal after range migration correction , the expression is as follows: ; S24: Azimuth compression.

[0009] Furthermore, the S24 azimuth compression includes: Azimuth matched filter is designed as follows: ; The echo signal after passing through the azimuth matched filter output is: ; Finally, the final compressed signal is obtained through azimuth IFFT is: ; Among them, is the amplitude of the azimuth impulse response; After range compression and azimuth compression imaging processing, the final target has been corrected to , and the target imaging processing is completed to obtain the SAR image.

[0010] Furthermore, the S5 includes: After object detection based on YOLO-TensorRT, the model outputs a sequence of prediction boxes , where B is the sequence of prediction boxes, is the i-th prediction box, and m is the total number of prediction boxes; , where and are the center point coordinates of the i-th prediction box, and are the width and height of the i-th prediction box respectively, is the confidence of the i-th prediction box, is the object category probability; Abnormal objects are eliminated through confidence filtering and geometric verification of the circumcircle to determine the final target, specifically: S51: Confidence filtering; Set a confidence threshold , which is used to filter out prediction boxes with higher confidence and obtain a valid prediction box set; the filtering conditions are as follows: ; in, It is a set of valid prediction boxes obtained after confidence screening; S52: Circumscribed circle calculation and verification; S521: Calculation of circumscribed circle; For the effective prediction box set Each valid prediction box in , calculate the coordinates of the center of its circumscribed circle and radius ; Since the effective prediction box is a rectangle, the center of its circumscribed circle is the center point of the rectangle, and the radius is the distance from the center point to the vertex of the rectangle, specifically: The calculation formula of the center coordinates is: ; Radius calculation formula: Assume that a vertex of the valid prediction box is , according to the distance formula between two points, the radius for: ; S522: Geometric verification of circumscribed circle; Detect the same target in N consecutive SAR images. Suppose the coordinates of the center of the circumscribed circle of the nth frame are , the radius is ; Calculate the mean of the center coordinates of N consecutive frames and the mean of the radius : ; Setting distance threshold and radius change threshold ; For each frame, the center coordinates and radius , and calculate its deviation from the mean: Center coordinate deviation : ; Radius deviation : ; like or , it is considered that the detection result of the frame is abnormal, and the valid prediction box corresponding to the SAR image of the frame is removed; S53: Final goal determined; After confidence screening and circumcircle geometric verification, the targets corresponding to the remaining valid prediction boxes are the final targets.

[0011] Further, the S6 includes: S61: Geocoding model; Let the longitude, latitude, and altitude obtained through GNSS measurement of the UAV platform be , and the error correction values of longitude, latitude, and altitude calculated by the ground base station be , respectively. Then the position of the UAV platform after differential correction is: ; where , , are the longitude, latitude, and altitude of the platform after differential correction, respectively; Combined with the SAR image, the circumcenter coordinates of the final target are used as the input for the conversion from pixel coordinates to geographic coordinates, and a mapping relationship from pixel coordinates to the geographic space is established; Assuming that the platform position is the origin, through the spherical trigonometry formula, the preliminary longitude and latitude of the target are calculated as: ; where and are the preliminary longitude increment and latitude increment, respectively; R e is the average radius of the earth; R is the slant range of the SAR from the target; φ is the azimuth angle of the target relative to the platform, determined according to the position of the pixel coordinates in the SAR image and the geometric relationship of SAR imaging; is the cosine value of the latitude of the UAV platform after differential correction; is the beam depression angle in SAR imaging; sin and cos are the sine function and cosine function, respectively; Then the preliminary longitude and latitude coordinates are: ; S62: Elevation correction; Fusing the digital elevation model data with a resolution of SRTM30m, the terrain undulation correction is performed on the preliminary longitude and latitude coordinates, specifically: Let the elevation value of the target area obtained from the SRTM data be , and the coordinate deviations caused by terrain undulation in the longitude and latitude directions be and , respectively; According to the geometric relationship, the coordinate deviation can be approximately calculated as: ; where tan is the tangent function; obtain the WGS-84 geodetic coordinates (lon, lat) of the final target: .

[0012] Furthermore, the S7 includes: S71: The edge computing platform transmits the WGS-84 geodetic coordinates of the final target to the ground station through the ground base station; after receiving the WGS-84 geodetic coordinates of the final target, the ground station generates the approaching flight path of the UAV through the path planning algorithm; S72: The UAV receives the approaching flight path instruction forwarded by the ground station through the ground base station, adjusts the heading and altitude, and flies towards the target area; the optical payload starts the high-definition imaging mode to verify the target details: The ground base station monitors the signal strength and the bit error rate BER of the communication link in real time; Let the signal strength threshold be , and the bit error rate threshold be ; When or , the ground base station dynamically switches the channel or increases the signal power.

[0013] The present invention also discloses a cooperative unmanned detection system based on SAR-guided optical payload, and the system includes: a UAV, an SAR, a data transmission communication radio, a voltage converter, an edge computing platform, a micro POS system, a power module, a radio frequency module, and an optical payload; The SAR is installed on the UAV, and the SAR includes an SAR antenna and an SAR host. The SAR antenna is used as the front-end component of the SAR for transmitting and receiving radar signals; the SAR host is composed of an FPGA main board, provides a gigabit Ethernet connection to the edge computing platform, integrates a radar waveform generation module, and supports dynamic adjustment of the pulse repetition frequency; the SAR processes the signal received by the SAR antenna 1 through the FPGA of the SAR host to obtain an SAR image; The data transmission communication radio is provided with a data transmission communication radio antenna for realizing data transmission between the SAR and the ground station; The voltage converter is located below the data transmission communication radio, and the voltage converter is connected to the data transmission communication radio, the SAR host, the FPGA, the edge computing platform, and the micro POS system through lines respectively for realizing high-low voltage conversion; The edge computing platform is connected to a switch through a gigabit Ethernet port, which is used to process optical payload images and SAR images in real time and supports the deployment of the YOLOv11-TensorRT acceleration model. The edge computing platform is connected to the data transmission communication radio through a line to achieve remote data transmission. The micro POS system is set near the SAR antenna. The micro POS system transmits the position and attitude information of the UAV to the edge computing platform in real time through an internal data bus to achieve the collaborative work of SAR and the optical payload 11. The power supply module consists of a battery and a voltage conversion module, which ensures the power supply requirements for the static state operation of the APOS in the case where the UAV does not provide power. In the case where the UAV provides power, it stabilizes and rectifies the input power to provide a unified secondary power supply for each module. The RF module consists of a transmitting channel, a receiving channel, and a frequency source. The transmitting channel performs up-conversion, filtering, and amplification on the digital intermediate frequency signal, enters the power amplifier module for power amplification, and then is sent to the transmitting antenna for external radiation. The receiving channel is responsible for amplifying, filtering, and gain controlling the echo signal received by the SAR antenna, and mixing it with the transmitting signal to obtain a difference frequency signal. The frequency source provides a coherent frequency reference source for each sub-unit of the SAR. The optical payload includes a dual-light pod, which is used for thermal image picture-in-picture switching, photographing and video recording, target tracking, and laser ranging.

[0014] Furthermore, the data transmission communication radio includes a control system, a digital-to-analog converter, an analog-to-digital converter, a timing and time synchronization module. The control system is used to complete the control of the functions of each module, including the gain control of the receiving channel, the working modes of the digital-to-analog converter and the analog-to-digital converter, and the real-time processing flow control. The analog-to-digital converter module completes the acquisition of the video signal of the receiving channel. The analog-to-digital converter module generates a digital broadband intermediate frequency signal. The timing module provides a timing reference for the digital-to-analog converter, the analog-to-digital converter, and the time synchronization module. The time synchronization module performs time synchronization on the pulse repetition frequency and processes real-time images.

[0015] The beneficial effects brought by the technical solution provided by the present invention at least include: Based on the heterogeneous computing architecture of FPGA and the edge computing platform, the present invention is equipped with hardware devices such as an SAR host, a POS, an SAR antenna, and an optoelectronic payload, and cooperates with a ground base station to achieve parallel processing of SAR real-time imaging processing, target detection, anomaly filtering, obtaining positioning information, realizing optoelectronic payload guidance control, and collaborative detection and recognition. Utilizing the wide-area detection ability of SAR that is all-weather, all-time, and not restricted by complex meteorological conditions and surface environments, it provides efficient target pre-search and area guidance for the optical payload, breaks through the limitations of the narrow field of view of the optical payload and its dependence on good lighting conditions, and realizes the integration of a wide-area scanning and close-range recognition unmanned system.

[0016] Through targeted demodulation processing of the echo signal, the present invention precisely optimizes the signal model, effectively strips redundant information, lays a high-precision data foundation for subsequent core signal processing processes such as range compression and azimuth processing, improves the accuracy of target position calculation, and further enhances the target recognition accuracy.

[0017] Through echo signal processing and heterogeneous computing architecture design, the present invention effectively avoids problems such as bandwidth competition between SAR massive data and flight control instructions and resource overload of a single hardware platform in traditional solutions, realizes reasonable splitting and efficient execution of data processing tasks, and ensures the stability and real-time processing ability of the system during high-load operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flowchart of the unmanned collaborative detection method based on SAR-guided optical payload provided by the present invention; Figure 2 is an overall structural schematic diagram of the unmanned collaborative detection system based on SAR-guided optical payload provided by the present invention; Figure 3 is a structural schematic diagram of the switch in the present invention; Figure 4 is a structural schematic diagram of the optical payload in the present invention; Figure 5 is a three-dimensional structural schematic diagram of the unmanned collaborative detection system based on SAR-guided optical payload provided by the present invention; Figure 6 is a schematic diagram of the UAV flight operation process; Figure 7 is the UAV flight route; Figure 8 is an imaging result diagram after range compression and azimuth compression processing; Figure 9 is the display of the SAR image detection result after deploying the vehicle; Figure 10 is the display of the pixel-geographic coordinate conversion result; Figure 11 is the display of the guidance and approach effect, where Figure 11 (a) is a schematic diagram of the actual target position; Figure 11 (b) is a first result schematic diagram of the guided optical payload towards the actual target position; Figure 11 (c) is a second result schematic diagram of the guided optical payload towards the actual target position.

[0019] Reference Signs: 1. SAR antenna; 2. Data transmission communication radio station; 3. Voltage converter; 4. SAR host; 5. Edge computing platform; 6. Miniature POS system; 7. Data transmission communication radio station antenna; 8. Switch; 9. Power supply module; 10. RF module; 11. Optical payload. Specific implementation manner

[0020] To make the objectives, technical solutions and advantages of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0021] Embodiment 1: This embodiment will describe the unmanned cooperative detection method based on SAR-guided optical payload of the present invention in more detail with reference to the corresponding drawings. As Figure 1 shown, it includes the following steps: S1: The SAR antenna 1 performs strip scanning on the target area to obtain an echo signal, and demodulates the echo signal to obtain the demodulated echo signal; S11: The SAR antenna 1 transmits a radar signal to the target area, and the expression of the echo signal reflected by the target and received is as follows: ; Among them, represents the echo signal received by the SAR after being reflected by the target; is the target reflection coefficient, which is a complex constant, t, respectively represent the slow-time dimension and fast-time dimension times, is the moment when the beam center passes through the target, and c is the electromagnetic wave propagation speed; represents the azimuth signal reception intensity, and a represents the azimuth direction, that is, the direction related to the SAR movement direction; represents the signal range-direction pulse envelope intensity; is the slant range of the SAR from the target at time t; exp() is the exponential function; is the center frequency, is the range-direction chirp rate; and respectively represent the signal range-direction pulse envelope function and the azimuth signal reception intensity function, and their expressions are as follows: ; Among them, is the pulse duration, P a is the azimuth antenna pattern function, rect() is the rectangular function, represents when ; sinc is the sinc function; is the angle between the target and the beam center line at time t; is the azimuth beam width, is the signal wavelength, is the azimuth antenna width; S12: Demodulate the echo signal to obtain the demodulated echo signal: ; wherein, represents the demodulated echo signal; j is the imaginary unit.

[0022] It should be noted that the Ku band adopted by the present invention has a working frequency of 12.5 - 18 GHz. This band has little ground interference, a high frequency, and is not easily affected by microwave radiation. The SAR system continuously emits radar signals to the target area. The echo signal reflected by the target and received by it, since in addition to the target information, the echo signal also includes carrier frequency information, it is necessary to remove the carrier frequency component in the echo signal through demodulation. The demodulated SAR signal also becomes the SAR raw data. The present invention performs targeted demodulation processing on the echo signal, precisely optimizes the signal model, effectively strips redundant information, lays a high-precision data foundation for subsequent core signal processing processes such as range compression and azimuth processing, improves the accuracy of target position calculation, and further enhances the target recognition accuracy.

[0023] S2: Input the demodulated echo signal into the FPGA, perform range and azimuth compression to achieve imaging, and obtain the SAR image; Input the demodulated echo signal as the SAR raw data into the FPGA, and perform range and azimuth compression to achieve imaging. The specific steps are as follows: S21: Range compression; Perform FFT on the range direction of the demodulated echo signal to obtain the following expression: ; wherein, is the echo signal after FFT processing in the range direction of the demodulated echo signal; G is the total gain including the scattering coefficient, set to 1; is the envelope of the range spectrum, is the range frequency; is the slant range between the SAR and the target at the slow time η; Range matching filter The expression of is: The echo signal after FFT processing passes through the range filter, outputs, and then passes through IFFT to obtain the expression after range compression: ; wherein, is the echo signal after range compression; is the inverse fast Fourier transform; is the IFFT expression of; S22: Azimuth FFT; The slant range of the SAR range target : , where is the initial slant range, v is the SAR platform velocity; Since the beam points close to the zero Doppler direction, and , approximate as a parabola, the slant range of the SAR range target , and the calculation formula can be replaced by: ; When performing azimuth FFT, the time-frequency relationship is , where is the azimuth frequency; is the azimuth chirp rate; is the SAR platform motion velocity; is the wavelength of the SAR transmitted signal; the azimuth FFT expression can be obtained as follows: ; where is the echo signal after azimuth FFT processing; is the fast Fourier transform of the time variable t; is the echo signal after range compression; is the azimuth frequency variable; is the frequency domain form of; is the range migration compensation term, and its expression is as follows: ; S23: Perform range migration correction; The range migration amount to be corrected , and the expression is as follows: ; The range migration correction factor , and the expression is as follows: ; Then the echo signal after range migration correction , and the expression is as follows: ; S24: Azimuth compression; The azimuth matched filter The design is as follows: ; The echo signal after passing through the azimuth matching filter output is: ; Finally, it passes through the azimuth IFFT to obtain the final compressed signal is: ; wherein, is the amplitude of the azimuth impulse response; After range compression and azimuth compression imaging processing, the final target has been corrected to and the target imaging processing is completed to obtain the SAR image.

[0024] It should be noted that after the echo signal received by the SAR antenna is sampled, it enters the FPGA as input data. These echo signals are usually wide pulse signals, containing the reflection information of targets at different ranges. Therefore, range compression and azimuth focusing are required to achieve imaging. The present invention uses range migration correction and azimuth pulse compression techniques to achieve range compression and azimuth focusing, and realizes real-time imaging.

[0025] S3: Transmit the SAR image to the edge computing platform through the COFDM data transmission link of the data transmission communication radio 2; It should be noted that through the design of echo signal processing and heterogeneous computing architecture (i.e., the FPGA and edge computing platform of the present invention) of the present invention, problems such as bandwidth competition between SAR massive data and flight control instructions and resource overload of a single hardware platform in traditional solutions are effectively avoided, and the reasonable splitting and efficient execution of data processing tasks are realized, ensuring the stability and real-time processing ability of the system under high-load operation.

[0026] S4: Perform INT8 quantization on the lightweight YOLOv11 model to obtain YOLO-TensorRT; perform target detection on the SAR image based on YOLO-TensorRT; determine whether a target is detected: when there is a target, perform S5; if no target is detected, return to S1; It should be further noted that in the target detection system of the present invention, we selected the lightweight YOLOv11 model and performed INT8 quantization on it. YOLOv11 was selected because it has high detection accuracy and fast inference speed in target detection tasks, and can meet the real-time requirements. And INT8 quantization is to further reduce the computational amount and storage requirements of the model and improve the inference efficiency.

[0027] During the inference process, the YOLOv11 model performs feature extraction and object detection on the input image through a series of operations such as convolution, pooling, and activation. Specifically, the model outputs multiple prediction boxes, each prediction box containing the bounding box information of the object (center point coordinates x, y, width w, height h), as well as the confidence conf of the object contained in the prediction box and the object class probability p class .

[0028] S5: Confidence screening and circumcircle geometric verification of object detection results to determine the final object; After object detection based on YOLO-TensorRT, the model outputs a sequence of prediction boxes , where B is the sequence of prediction boxes, is the i-th prediction box, and m is the total number of prediction boxes; , where and are the center point coordinates of the i-th prediction box, and are the width and height of the i-th prediction box respectively, is the confidence of the i-th prediction box, is the object class probability; Abnormal objects are removed through confidence screening and circumcircle geometric verification to determine the final object, specifically: S51: Confidence screening; Set a confidence threshold to filter out prediction boxes with higher confidence to obtain a set of valid prediction boxes; the screening conditions are as follows: ; Among them, is the set of valid prediction boxes obtained after confidence screening; through this screening process, prediction boxes with lower confidence and likely to be misdetected can be removed; S52: Circumcircle calculation and verification; S521: Circumcircle calculation; For each valid prediction box in the set of valid prediction boxes , calculate the center coordinates and radius of its circumcircle; since the valid prediction box is a rectangle, the center of its circumcircle is the center point of the rectangle, and the radius is the distance from the center point to the vertex of the rectangle, specifically: Center coordinate calculation formula: ; Radius calculation formula: Assume that a vertex of the valid prediction box is , according to the distance formula between two points, the radius is: ; S522: Geometric verification of the circumcircle; Detect consecutive N frames of SAR images of the same target. Let the coordinates of the center of the circumcircle of the nth frame be and the radius be ; Calculate the mean value of the center coordinates of consecutive N frames and the mean value of the radius : ; Set the distance threshold and the radius change threshold ; For the center coordinates and the radius of each frame, calculate their deviations from the mean value: Deviation of center coordinates : ; Deviation of radius : ; If or , it is considered that the detection result of this frame is abnormal, and the effective prediction box corresponding to this frame of SAR image is removed; S53: Final target determination; After confidence screening and circumcircle geometric verification, the target corresponding to the remaining effective prediction boxes is the final target.

[0029] It should be noted that through confidence screening and geometric verification related to the circumcircle, abnormal targets can be effectively removed, improving the accuracy and reliability of the target detection result, and providing more stable and accurate target information for subsequent processing steps.

[0030] S6: Combine the SAR image, use the coordinates of the center of the circumcircle of the final target as the input for pixel coordinate to geographic coordinate conversion, perform pixel coordinate to geographic coordinate conversion, and perform elevation correction through digital elevation model data to obtain the WGS-84 geographic coordinates of the final target; S61: Geocoding model The GNSS / IMU data of the system combined with the differential GNSS data provided by the ground base station can provide longitude and latitude information. The ground base station continuously monitors satellite signals, calculates the error correction value of satellite signals in this area, and broadcasts it.

[0031] Let the longitude, latitude, and altitude measured by the GNSS of the UAV platform be , the error correction values of longitude, latitude, and altitude calculated by the ground base station are respectively , then the position of the UAV platform after differential correction is: ; where , , are respectively the longitude, latitude, and altitude of the platform after differential correction; Combined with the SAR image, the circumcenter coordinates of the final target are used as the input for the conversion from pixel coordinates to geographic coordinates, and a mapping relationship from pixel coordinates to the geographic space is established; Assuming that the platform position is the origin, through the spherical trigonometric formula, the preliminary longitude and latitude of the target are calculated: ; where and are respectively the preliminary longitude increment and latitude increment; R e is the average radius of the earth; R is the slant range of the SAR from the target; φ is the azimuth angle of the target relative to the platform, which is determined according to the position of the pixel coordinates in the SAR image and the geometric relationship of SAR imaging; is the cosine value of the latitude of the UAV platform after differential correction; is the beam depression angle in SAR imaging; sin and cos are the sine function and cosine function respectively; Then the preliminary longitude and latitude coordinates are: ; S62: Elevation correction; Fusing the digital elevation model data with a resolution of 30 m of SRTM, the terrain undulation correction is performed on the preliminary longitude and latitude coordinates. Let the elevation value of the target area obtained from the SRTM data be , and the coordinate deviations caused by terrain undulation in the longitude and latitude directions are respectively and .

[0032] According to the geometric relationship, the coordinate deviation can be approximately calculated: ; Finally, the accurate WGS-84 geographic coordinates (lon, lat) are obtained: .

[0033] S7: The edge computing platform 5 transmits the WGS-84 geographical coordinates of the final target to the ground station through the ground base station. The UAV receives the approaching flight path instruction forwarded by the ground station through the ground base station, adjusts its heading and altitude, and flies towards the target area; the optical payload activates the high-definition imaging mode to verify the target details. S71: Guidance instruction generation; The edge computing platform 5 transmits the converted target coordinates (such as in JSON format {"lon":116.4034,"lat":39.9234,"radius":5.0}) to the ground station through the ground base station. The ground base station plays a key communication hub role during data transmission. It uses an anti-interference digital link (such as COFDM) to ensure stable data transmission in a complex electromagnetic environment. After receiving the WGS-84 geographical coordinates of the final target, the ground station combines the preset safety radius (r is the target feature radius), and generates the approaching flight path of the UAV through a path planning algorithm (such as algorithm). In the algorithm, the cost function f(e)=g(e)+h(e) of each node is calculated, where g(e) is the actual cost from the starting node to the current node e, and h(e) is the heuristic estimated cost from the current node e to the target node. Let the starting node be S, the target node be M, and the current node be e. The distance between any two nodes can be calculated based on the longitude and latitude coordinates: ; where, (lon1, lat1) are the longitude and latitude coordinates of node e1; (lon2, lat2) are the longitude and latitude coordinates of node e2, and e1 and e2 are used to represent any two nodes; Then it can be obtained that g(e) is the sum of the distances of all edges passed from the starting node S to node e; h(e) uses the straight-line distance as the heuristic estimate: h(e)=d(e, M); S72: Task closed-loop; The UAV receives the instruction forwarded by the ground station through the ground base station, adjusts its heading and altitude, and flies towards the target area. The optical payload activates the high-definition imaging mode (such as 30x zoom) to verify the target details.

[0034] The ground base station also plays a role of dynamic compensation throughout the process. Due to the complex electromagnetic environment, which may cause attenuation and interference of communication signals, the ground base station monitors the communication link quality in real time, such as signal strength and bit error rate BER and other indicators.

[0035] Let the signal strength threshold be , the bit error rate threshold is . When or , the ground base station dynamically switches channels or enhances the signal power. For example, by adjusting the gain of the transmitting antenna to enhance the signal power: ; where is the original transmitting power, and by dynamically adjusting to ensure the stability of the communication link, thereby ensuring the stability of the command transmission and the pod's backhaul image, forming a complete task closed-loop of "discovery - guidance - verification".

[0036] Embodiment 2: This embodiment will describe the unmanned cooperative detection system based on SAR-guided optical payload of the present invention in more detail with corresponding drawings. As Figure 1 shown, the system includes: an unmanned aerial vehicle, SAR, a digital data communication radio 2, a voltage converter 3, an edge computing platform 5, a micro POS system 6, a power module 9, a radio frequency module 10, and an optical payload 11; The SAR is installed on the unmanned aerial vehicle. The SAR includes a SAR antenna 1 and a SAR host 4. The SAR antenna 1, as the front-end component of the SAR, is used for the transmission and reception of radar signals. The SAR host 4 is composed of an FPGA main board, provides a gigabit Ethernet connection to the edge computing platform 5, integrates a radar waveform generation module, and supports dynamic adjustment of the pulse repetition frequency. The SAR processes the signals received by the SAR antenna 1 through the FPGA of the SAR host to obtain a SAR image. Further explanation, the SAR antenna of the present invention has high gain and directional beam characteristics, operates in the X-band (8 - 12 GHz), the gain ≥ 25 dBi, horizontal polarization mode, is installed outside the unmanned aerial vehicle pod, and uses lightweight composite materials to reduce wind resistance. The SAR host in the present invention, the core signal processing unit, uses an FPGA main board Xilinx Kintex-7 XC7K325T, and can complete pulse compression, MTI filtering, and data scheduling.

[0037] The digital data communication radio 2 is provided with a digital data communication radio antenna 7 for realizing data transmission between the SAR and the ground station. Further explanation, the digital data communication radio is located on the right side of the integrated platform, and its position is conducive to the transmission and reception of wireless signals and reduces interference from internal modules. The digital data communication radio 2 supports the L-band, the maximum transmission distance is 50 km, the modulation method is QPSK, and it has a built-in encryption module and supports RS422 / Ethernet interfaces.

[0038] The voltage converter 3 is located below the data transmission communication radio 2, and the voltage converter 3 is connected to the data transmission communication radio 2, the SAR host 4, the FPGA, the edge computing platform, and the micro POS system 6 through lines respectively, for realizing high-voltage and low-voltage conversion. Further explanation, the voltage converter is close to the power supply module, which is convenient for processing the input power supply, stabilizing the input power supply, and ensuring the stable operation of each module.

[0039] The edge computing platform 5 is connected to the switch 8 through a gigabit network port, for processing the optical payload image and the SAR image in real time, and supporting the deployment of the YOLOv11-TensorRT acceleration model; the edge computing platform 5 is connected to the data transmission communication radio 2 through a line to realize remote data transmission. It should be noted that the edge computing platform (NVIDIA Jetson Orin NX) in the present invention supports channel pruning + INT8 quantization.

[0040] The micro POS system 6 is arranged close to the SAR antenna 1. The micro POS system 6 transmits the position and attitude information of the unmanned aerial vehicle to the edge computing platform 5 in real time through an internal data bus, to realize the collaborative work of the SAR and the optical payload 11; it should be noted that the micro POS system (positioning and orientation system) integrates modules such as GNSS, IMU, and inertial navigation. By fusing the data of these components, the position and attitude information related to the radar signal can be obtained more accurately. GNSS (Global Navigation Satellite System) is mainly used to provide positioning information, providing basic data for the system to determine the geographical location of the target. The IMU (Inertial Measurement Unit) is used to measure attitude angles and other information. In this embodiment, the parameter settings of the micro POS system are as follows: the GNSS positioning accuracy is ±0.1 m, the attitude angle accuracy is ±0.1, the output frequency is 100 Hz; it integrates IMU, inertial navigation, etc., and is synchronized with the radar PRF through the PPS signal. In addition, the micro POS system is used to obtain the position and attitude information related to the radar signal, providing an accurate spatio-temporal reference for subsequent processing.

[0041] The power supply module 9 is composed of a battery and a voltage conversion module, ensuring the power supply requirement for the static state operation of the APOS (Airborne Position and Orientation System) airborne position and attitude system in the case that the unmanned aerial vehicle does not provide power; in the case that the unmanned aerial vehicle provides power, it stabilizes and rectifies the input power supply, providing a unified secondary power supply for each module; after the voltage converter 3 stabilizes and rectifies the input power supply, it is connected to modules such as the edge computing platform 5, the data transmission communication radio 2, and the micro POS system 6 through lines, ensuring the stable operation of each module.

[0042] The radio frequency module 10 consists of a transmitting channel, a receiving channel, and a frequency source. The transmitting channel performs up-conversion, filtering, and amplification on the digital intermediate frequency signal, enters the power amplifier module for power amplification, and then is sent to the transmitting antenna for external radiation. The receiving channel is responsible for amplifying, filtering, and gain control of the echo signal of the receiving antenna, and mixing it with the transmitting signal to obtain a difference frequency signal. The frequency source provides a coherent frequency reference source for each sub-unit of the SAR. The optical payload 11 includes a dual-optical pod, which is used for thermal image picture-in-picture switching, photographing and video recording, target tracking, and laser ranging. Further, in this embodiment, the pod uses Q30TIRMplus, which is a high-precision three-axis stabilized dual-optical pod equipped with a 30x optical zoom Sony camera with 2.13 million pixels, a thermal imager with a 25 mm lens and a resolution of 640x480, and a 2000-meter rangefinder. Among them, the dual-optical pod supports visible light zooming, thermal image picture-in-picture switching, multi-color plate switching, photographing and video recording, target tracking function, thermal image digital zooming, laser ranging, etc.

[0043] Furthermore, the data transmission communication radio 2 includes a control system, an analog-to-digital converter (ADC), a digital-to-analog converter (DAC), timing and time synchronization, and a recording module. The control system is used to complete the control of the functions of each module, including the gain control of the receiving channel, the working modes of the ADC and DAC, and the real-time processing flow control. The ADC module completes the acquisition of the video signal of the receiving channel, and the DAC module generates a digital broadband intermediate frequency signal. The timing module provides a timing reference for the ADC, DAC, and time synchronization module, and the time synchronization module performs time synchronization on the pulse repetition frequency and processes real-time images.

[0044] It should be noted that the present invention is based on a heterogeneous computing architecture of FPGA and an edge computing platform, equipped with hardware devices such as a SAR host, a POS, a SAR antenna, and an optoelectronic payload, and cooperates with a ground base station to achieve parallel processing of SAR real-time imaging processing, target detection, anomaly filtering, obtaining positioning information, realizing optoelectronic payload guidance control, and collaborative detection and recognition. Utilizing the wide-area detection ability of SAR that is all-weather, all-time, and not restricted by complex meteorological conditions and surface environments, it provides efficient target pre-search and area guidance for the optical payload, breaks through the limitations of the narrow field of view of the optical payload and its dependence on good lighting conditions, and realizes the integration of a wide-area scanning and close-range recognition unmanned system.

[0045] Simulation experiment The effectiveness of the method proposed by the present invention is evaluated through a simulation scenario. This experiment is based on Aerospace Macro Figure SixBuild a test platform for the rotary-wing UAV. The ground-end remote control device and computer two form the takeoff and flight path controller of the UAV, and the base station is used to record the flight path trajectory of the UAV. After the UAV takes off and stably enters the flight path, computer one can issue relevant commands through the data transmission link of the ground radio station and the air radio station to control the power on and off of the SAR host to collect SAR raw data and perform subsequent data processing steps.

[0046] A schematic diagram of the UAV flight operation process is as Figure 6 shown. First, turn on the SAR system and scan the selected area in a strip scan mode. Adopting this method can search a large area quickly, preliminarily determine the approximate location of the suspicious (interesting) target, and transmit the location information to the pod to guide it towards the specified direction for further detailed observation.

[0047] SAR antenna strip scan The UAV flies according to the predetermined flight path, and the flight path is as Figure 7 shown. The UAV vertically rises from point A. After the flight altitude reaches 250 meters, it enters the flight path, turns on the SAR system switch, and the working mode is strip scan, and the right-looking method is adopted. Before the UAV is about to reach each corner, it needs to turn off the SAR system switch first, and then turn on the switch until it stably enters the next straight line. When the UAV flies to point B, the flight task is completed.

[0048] FPGA real-time imaging processing Input the collected raw SAR data into the FPGA, perform range-direction pulse compression and azimuth-direction pulse compression processing, and image to obtain the imaging result as Figure 8 shown.

[0049] Data link transmission The SAR image obtained by FPGA imaging processing is transmitted to the edge terminal platform (Jetson Orin) in real time through the COFDM data transmission link in the form of SAR image data blocks (JPEG format).

[0050] Real-time object detection based on YOLO-TensorRT The SAR image input to the edge terminal is processed by the YOLO-TensorRT algorithm to obtain the result after algorithm processing.

[0051] Confidence screening and geometric verification As Figure 9 shown, a series of target prediction boxes can be obtained after object detection. After these prediction boxes are processed by confidence screening and circumscribed circle geometric verification, the final target prediction boxes can be screened out.

[0052] Pixel to geographic coordinate conversion As Figure 10 shown, after confidence screening and circumcircle stability verification, the targets corresponding to the retained prediction boxes are the finally determined valid targets. The circumcenter coordinates of the circumcircles of these targets will be used as the input for the subsequent conversion from pixel coordinates to geographic coordinates, so as to obtain the actual geographic coordinate positions of the targets.

[0053] Optical Payload Guidance and Proximity Flight As Figure 11 shown, after obtaining the actual geographical positions of the targets, as shown in Figure 11 (a), guide the pod towards the designated position, as shown in Figure 11 (b) and Figure 11 (c), and obtain more detailed information about the targets, as shown in Figure 11 shown.

[0054] It should be noted that the serial numbers of the embodiments of the present invention above are only for description and do not represent the advantages or disadvantages of the embodiments. And the terms "including", "comprising" or any other variant thereof in this text are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, device, article or method. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of another identical element in the process, device, article or method including the element.

Claims

1. An unmanned collaborative detection method based on an SAR-guided optical payload, characterized in that, Including: S1: Perform strip scanning on the target area through the SAR antenna to obtain the echo signal, and demodulate the echo signal to obtain the demodulated echo signal; S2: Input the demodulated echo signal into the FPGA, perform range and azimuth compression to achieve imaging, and obtain the SAR image; S3: Transmit the SAR image to the edge computing platform through the COFDM data transmission link of the data transmission communication radio; S4: Perform INT8 quantization on the lightweight YOLOv11 model to obtain YOLO-TensorRT; Perform target detection on the SAR image based on YOLO-TensorRT; Determine whether a target is detected: When a target exists, perform S5; If no target is detected, return to S1; S5: Screen the confidence level of the detection result of target detection and perform geometric verification of the circumscribed circle to determine the final target; S6: Combine the SAR image, use the center coordinates of the circumscribed circle of the final target as the input for the conversion from pixel coordinates to geographic coordinates, perform the conversion from pixel coordinates to geographic coordinates, and perform elevation correction through digital elevation model data to obtain the WGS-84 geographic coordinates of the final target; S7: The edge computing platform transmits the WGS-84 geographic coordinates of the final target to the ground station through the ground base station. The unmanned aerial vehicle receives the approach flight path instruction forwarded by the ground station through the ground base station, adjusts the heading and altitude, and flies towards the target area; The optical payload activates the high-definition imaging mode to verify the target details.

2. The unmanned collaborative detection method based on the SAR-guided optical payload according to claim 1, characterized in that, The said S1 includes: S11: Transmit a radar signal to the target area through the SAR antenna. The expression of the echo signal after being reflected by the target and received is as follows: ； Among them, represents the echo signal received by the SAR after being reflected by the target; is the target reflection coefficient, which is a complex constant. t and represent the slow-time dimension and the fast-time dimension of time respectively, is the moment when the beam center passes through the target, and c is the propagation speed of electromagnetic waves; represents the received intensity of the azimuth signal. a represents the azimuth direction, that is, the direction related to the movement direction of the SAR; represents the pulse envelope intensity of the signal in the range direction; is the slant range between the SAR and the target at time t; exp() is the exponential function; is the center frequency, is the frequency modulation rate in the range direction; and represent the pulse envelope function of the signal in the range direction and the received intensity function of the azimuth signal respectively, and their expressions are as follows: ； Among them, is the pulse duration, P a is the azimuth antenna pattern function, rect() is the rectangle function, means when ; sinc is the sinc function; is the angle between the target and the beam centerline at time t; is the azimuth beam width, is the signal wavelength, is the azimuth antenna width; S12: Demodulate the echo signal to obtain the demodulated echo signal: ； Among them, represents the demodulated echo signal; j is the imaginary unit.

3. The unmanned collaborative detection method based on SAR-guided optical payload according to claim 2, characterized in that The said S2 includes: Input the demodulated echo signal as the SAR raw data into the FPGA, perform range and azimuth compression to achieve imaging. The specific steps are as follows: S21: Range compression; Perform FFT on the range direction of the demodulated echo signal to obtain the following expression: ； Among them, is the echo signal after FFT processing in the range direction of the demodulated echo signal; G is the total gain including the scattering coefficient, set to 1; is the envelope of the range spectrum, is the range frequency; is the slant range between the SAR and the target at the slow time η; Range-matching filter The expression is as follows: ； The echo signal after FFT processing is output through a range filter and then through IFFT to obtain the expression after range compression. : ； Among them, is the inverse fast Fourier transform; is the IFFT expression of S22: Azimuth FFT; SAR slant range to the target : ； wherein, is the initial slant range, and v is the SAR platform velocity; Since the beam direction is close to the zero Doppler direction, and , the is approximated as a parabola, and the slant range of the SAR from the target , and the calculation formula can be replaced with: ； When performing azimuth FFT, the time-frequency relationship is , where is the azimuth frequency; is the azimuth chirp rate; is the SAR platform moving speed; is the wavelength of the SAR transmitted signal. The azimuth FFT expression can be obtained as follows: ； Among them, is the echo signal after FFT processing in the azimuth direction; is the fast Fourier transform of the time variable t; is the echo signal after range compression; is the azimuth frequency variable; is in the frequency domain form; is the range migration compensation term, and its expression is as follows: ； S23: Perform range migration correction; Range migration amount to be corrected , and the expression is as follows: ； Range migration correction factor , and the expression is as follows: ； The echo signal after range migration correction , and the expression is as follows: ； S24: Azimuth compression.

4. The unmanned collaborative detection method based on the SAR-guided optical payload according to claim 3, wherein, The said S24 azimuth pulse compression includes: Azimuth matching filter The design is as follows: ； The echo signal after passing through the azimuth matching filter output is as follows: ； Finally, the azimuth IFFT is performed to obtain the final compressed signal which is ； Among them, is the amplitude of the azimuth impulse response; After range compression and azimuth compression imaging processing, the final target has been corrected to and the target imaging processing is completed to obtain the SAR image.

5. The method for unmanned collaborative detection based on an SAR-guided optical payload according to claim 4, wherein The said S5 includes: After object detection based on YOLO-TensorRT, the model outputs a sequence of predicted bounding boxes , where B is the sequence of predicted bounding boxes, is the i-th predicted bounding box, and m is the total number of predicted bounding boxes; , where and are the center coordinates of the i-th predicted bounding box, and are the width and height of the i-th predicted bounding box respectively, is the confidence of the i-th predicted bounding box, is the probability of the target category; Eliminate abnormal targets through confidence level screening and geometric verification of the circumscribed circle to determine the final target. Specifically: S51: Confidence level screening; Set a confidence threshold , which is used to filter out the prediction boxes with higher confidence to obtain an effective prediction box set; the filtering conditions are as follows: ； Among them, is the set of effective prediction bounding boxes obtained after confidence filtering; S52: Circumscribed circle calculation and verification; S521: Circumscribed circle calculation; For the set of valid prediction boxes For each valid prediction box in it, calculate the center coordinates and radius of its circumscribed circle; Since the valid prediction box is a rectangle, the center of its circumscribed circle is the center point of the rectangle, and the radius is the distance from the center point to the vertex of the rectangle. Specifically: Calculation formula for the center coordinates: ； Calculation formula for the radius: Suppose a vertex of the effective prediction box is , according to the distance formula between two points, the radius is: ； S522: Geometric verification of the circumscribed circle; Detect consecutive N-frame SAR images of the same target. Let the coordinates of the center of the circumcircle of the nth frame be , and the radius be ; Calculate the mean of the center coordinates of consecutive N frames and the mean of the radii : ； Set distance threshold and radius change threshold ; For the center coordinates and radius of each frame, calculate the deviation from the mean: Deviation of center coordinates : ； Radius deviation : ； If or , it is considered that there is an abnormality in the detection result of this frame, and the valid prediction box corresponding to the SAR image of this frame is removed; S53: Final target determination; After confidence level screening and geometric verification of the circumscribed circle, the target corresponding to the remaining valid prediction boxes is the final target.

6. The unmanned collaborative detection method based on an SAR-guided optical payload according to claim 5, characterized in that The said S6 includes: S61: Geocoding model; Let the longitude, latitude, and altitude obtained by GNSS measurement through the drone platform be , and the error correction values of longitude, latitude, and altitude calculated by the ground base station be , respectively. Then the position of the drone platform after differential correction is: ； Among them, , , are the longitude, latitude and altitude of the platform after differential correction, respectively; Combined with the SAR image, use the center coordinates of the circumcircle of the final target as the input for the conversion from pixel coordinates to geographic coordinates, and establish the mapping relationship from pixel coordinates to the geospatial space; Assume that the platform location is the origin, and the preliminary longitude and latitude of the target are calculated through spherical trigonometric formulas : ； Among them, and are the preliminary longitude increment and latitude increment respectively; R e is the average radius of the earth; R is the slant range of the SAR distance target; φ is the azimuth angle of the target relative to the platform, which is determined according to the position of the pixel coordinates in the SAR image and the geometric relationship of SAR imaging; is the cosine value of the latitude of the UAV platform after differential correction; is the beam depression angle in SAR imaging; sin and cos are the sine function and cosine function respectively; Then the preliminary longitude and latitude coordinates are: ； S62: Elevation correction; Fuse the digital elevation model data with a resolution of SRTM30m to correct the terrain undulation of the preliminary longitude and latitude coordinates. Specifically: Let the elevation value of the target area obtained from SRTM data be , and the coordinate deviations caused by terrain undulation are respectively and ; According to the geometric relationship, the coordinate deviation can be approximately calculated: ； Where, tan is the tangent function; Obtain the WGS-84 geographic coordinates (lon, lat) of the final target: 。 7. The unmanned collaborative detection method based on SAR-guided optical payload according to claim 6, characterized in that, The said S7 includes: S71: The edge computing platform transmits the WGS-84 geographical coordinates of the final target to the ground station through a ground base station; after receiving the WGS-84 geographical coordinates of the final target, the ground station generates a close-proximity flight path for the UAV through a path planning algorithm; S72: The UAV receives the close-proximity flight path instruction forwarded by the ground station through the ground base station, adjusts its heading and altitude, and flies towards the target area; the optical payload activates the high-definition imaging mode to verify the target details: The ground base station monitors the signal strength and bit error rate BER of the communication link in real time and the bit error rate BER; Set the signal strength threshold as , and the bit error rate threshold as ; When or occurs, the ground base station dynamically switches channels or enhances the signal power.

8. An unmanned cooperative detection system based on an SAR-guided optical payload, for implementing the unmanned cooperative detection method based on an SAR-guided optical payload according to any one of claims 1-7, characterized in that, The system includes: a UAV, an SAR, a data transmission communication radio, a voltage converter, an edge computing platform, a micro POS system, a power supply module, a radio frequency module, and an optical payload; The SAR is installed on the UAV. The SAR includes an SAR antenna and an SAR host. The SAR antenna, as the front-end component of the SAR, is used for transmitting and receiving radar signals; the SAR host is composed of an FPGA main board, provides a gigabit Ethernet connection to the edge computing platform, integrates a radar waveform generation module, and supports dynamic adjustment of the pulse repetition frequency; the SAR processes the signals received by the SAR antenna through the FPGA of the SAR host to obtain an SAR image; The data transmission communication radio is equipped with a data transmission communication radio antenna for realizing data transmission between the SAR and the ground station; The voltage converter is located below the data transmission communication radio, and the voltage converter is connected to the data transmission communication radio, the SAR host, the FPGA, the edge computing platform, and the micro POS system through lines respectively for realizing high-voltage and low-voltage conversion; The edge computing platform is connected to a switch through a gigabit network port for real-time processing of optical payload images and SAR images, and supports the deployment of the YOLOv11-TensorRT acceleration model; the edge computing platform is connected to the data transmission communication radio through a line to realize remote data transmission; The micro POS system is arranged close to the SAR antenna. The micro POS system transmits the position and attitude information of the UAV to the edge computing platform in real time through an internal data bus to realize the collaborative work of the SAR and the optical payload; The power supply module consists of a battery and a voltage conversion module, which ensures the power supply requirements for the stationary state operation of the APOS in the case where the UAV does not provide power; in the case where the UAV provides power, it stabilizes and rectifies the input power to provide a unified secondary power supply for each module; The radio frequency module consists of a transmitting channel, a receiving channel, and a frequency source; the transmitting channel performs up-conversion, filtering, and amplification processing on the digital intermediate frequency signal, enters the power amplifier module for power amplification, and then is sent to the transmitting antenna for external radiation; the receiving channel is responsible for amplifying, filtering, and gain control of the echo signal received by the receiving SAR antenna, and mixing with the transmitting signal to obtain a difference frequency signal; the frequency source provides a coherent frequency reference source for each sub-unit of the SAR; The optical payload includes a dual-optical pod for thermal image picture-in-picture switching, photographing and video recording, target tracking, and laser ranging.

9. The unmanned cooperative detection system based on the SAR-guided optical payload according to claim 8, characterized in that, The data transmission communication radio station includes a control system, a digital-to-analog converter, an analog-to-digital converter, and a timing and time synchronization module; the control system is used to control the functions of each module, including the gain control of the receiving channel, the working modes of the digital-to-analog converter and the analog-to-digital converter, and the real-time processing flow control; the analog-to-digital converter module completes the acquisition of the video signal of the receiving channel. The analog-to-digital converter module generates a digital broadband intermediate frequency signal; the timing module provides a timing reference for the digital-to-analog converter, the analog-to-digital converter, and the time synchronization module; the time synchronization module performs time synchronization on the pulse repetition frequency and processes real-time images.

Images