Target stable tracking method combined with ground velocity compensation

Abstract

The invention relates to the technical field of unmanned aerial vehicle target tracking, in particular to a target stable tracking method combined with ground velocity compensation. The method includes measuring flight speed and direction of the unmanned aerial vehicle and determining ground speed; acquiring an angle parameter on the flight path of the unmanned aerial vehicle by using an angle measurement system; calculating a ground speed compensation value according to the ground speed and the angle parameter; the photoelectric turret captures a visible light image and an infrared image and carries out real-time registration fusion; extracting local contrast and information entropy according to the registered and fused image, and dynamically adjusting the fusion proportion of the visible light image and the infrared image through a fusion tracking method to generate a fusion tracking image; inputting the fused tracking image into a related tracking algorithm, and performing target tracking; performing target search area offset according to the ground velocity compensation parameters; and the servo system inputs the ground speed compensation parameter into a speed closed-loop control loop, and stably tracks the target in combination with the miss distance. The method has the advantages that real-time registration is achieved, and the image registration efficiency is improved; and the target tracking performance is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle target tracking, and in particular to a method for stable target tracking combined with ground speed compensation. Background Art

[0002] In the field of UAV airborne optoelectronic turret technology, target tracking technology is a core component, which directly affects the efficiency and accuracy of reconnaissance, surveillance and target identification tasks. With the development of UAV technology, higher requirements are put forward for target tracking technology, especially the stability and accuracy in dynamic flight environments. In the UAV airborne optoelectronic turret system, achieving stable target tracking is crucial to improving the efficiency and effectiveness of UAV mission execution.

[0003] In the prior art, the optoelectronic turret system onboard a UAV usually includes a visible light camera and an infrared camera, which can provide visible light images and infrared images of the target. However, due to the influence of flight speed and attitude changes during high-speed or maneuvering flight, the computing power of the onboard embedded computer is limited, and there are certain limitations in target tracking. Especially in a dynamically changing environment, the UAV needs to adjust the direction of its optoelectronic turret in real time to maintain stable tracking of the target, which is very challenging technically.

[0004] First, when a drone is flying at high speed or performing maneuvering flight, the target tracking system has difficulty accurately predicting the target's movement trajectory due to changes in flight speed and attitude. Second, existing systems often lack an effective ground speed compensation mechanism and cannot accurately compensate for the drone's movement relative to the ground, thus affecting tracking stability. In addition, the image registration algorithm of the existing technology usually cannot achieve real-time processing, which limits the real-time and reliability of target tracking.

[0005] In practical applications, the optoelectronic turret system onboard a UAV needs to perform tasks under various environmental conditions, including different flight altitudes, speeds, and complex weather conditions. These factors pose challenges to the performance of the target tracking system. For example, when flying at high speeds, the UAV needs to quickly adjust the direction of the optoelectronic turret to maintain stable tracking of the target; in complex weather conditions, such as fog, rain, or at night, the UAV needs to use multimodal information such as infrared images to track the target.

[0006] In order to improve the accuracy and stability of target tracking, researchers have proposed a variety of methods, including model-based prediction methods, image processing techniques, and machine learning algorithms, etc. However, these methods still have shortcomings in terms of real-time performance, accuracy, and robustness, especially in the dynamic flight environment of UAVs.

[0007] In addition, the image registration algorithms in the prior art usually require a complex ground calibration process, including accurate measurement of the camera's visual axis center offset, edge distortion, and pixel mapping relationship. These calibration processes are not only time-consuming and labor-intensive, but also difficult to adapt to the application requirements of zoom cameras in the optoelectronic turret system of drones.

[0008] In summary, the existing UAV-mounted optoelectronic turret system faces problems such as poor adaptability to dynamic flight environments, insufficient ground speed compensation, and poor real-time image registration in terms of target tracking. These problems limit the efficiency and effectiveness of UAVs in performing tasks in complex environments. Therefore, it is necessary to develop a new target stabilization tracking method combined with ground speed compensation to improve the tracking performance of UAVs in dynamic flight environments. Summary of the invention

[0009] In order to solve the above problems, the present invention provides a method for stable target tracking combined with ground speed compensation.

[0010] The present invention aims to provide a method for stable target tracking combined with ground speed compensation, which specifically comprises the following steps: S1. Measure the flight speed and direction of the drone through the drone-mounted IMU to determine the ground speed; use the angle measurement system of the drone-mounted optoelectronic turret to obtain the angle parameters on the drone's flight path; calculate the ground speed compensation value based on the ground speed and angle parameters; S2. Install both the visible light camera and the infrared camera on the optoelectronic turret on the drone, use the optoelectronic turret on the drone to capture visible light images and infrared images, and perform real-time registration and fusion through image registration algorithm; S3. Extract local contrast and information entropy based on the registered fused image, and dynamically adjust the fusion ratio of the visible light image and the infrared image through the fusion tracking method to generate a fused tracking image; input the fused tracking image into the relevant tracking algorithm to track the target; offset the target search area based on the parameters of the ground speed compensation to ensure that the target is searched with the maximum probability; S4. The servo system inputs the ground speed compensation parameter into the speed closed-loop control loop, and combines it with the miss distance of the relevant tracking to stably track the target.

[0011] Preferably, the angle parameters in step S1 include an azimuth angle α and a pitch angle β; The method for calculating the ground speed compensation value specifically includes: S101. Define the eastward speed of the drone by the azimuth angle α and northbound speed ; Set the drone's eastward speed and northbound speed Converted to velocity component relative to the UAV flight path : ; In the formula, α represents the azimuth angle of the UAV flight path, represents the eastward speed, represents the north speed; S102. Adjust the velocity component according to the pitch angle β on the UAV flight path , and the adjusted horizontal velocity component is obtained, and the calculation formula is as follows: ; In the formula, Based on the eastward speed of the drone and northbound speed , the horizontal velocity component calculated from the azimuth angle α and the pitch angle β; represents the velocity component; β represents the pitch angle on the flight path of the UAV; S103. Calculate the ground speed compensation value; the calculation formula is as follows: ; In the formula, Indicates the ground speed compensation value; is a proportionality constant used to adjust the compensation intensity; Represents the horizontal velocity component.

[0012] Preferably, step S2 specifically includes the following sub-steps: S201. Ground calibration: calibrate the visible light Y-direction field angle of the visible light camera at each focal length and the infrared Y-direction field angle of the infrared camera at each focal length on the ground, record the relationship between the field angle and the focal length as a file, and store it on the camera control board; S202. Coarse matching of field of view angle: The camera control board controls the movement of the visible light camera or the infrared camera through the PID algorithm so that the field of view angle of a single pixel of the visible light image is equal to the field of view angle of a single pixel of the infrared image; S203. Image descriptor extraction and registration: extract heterogeneous image descriptors of the visible light image and the infrared image, match the descriptors, store the projection relationship of each pixel coordinate of the visible light image and the infrared image as matrix A; call matrix A to complete the registration.

[0013] Preferably, in step S203, the heterogeneous image descriptors of the visible light image and the infrared image include a phase consistency descriptor and a gradient descriptor; and the weighted correlation distance method is used for matching, and the specific method is as follows: S2031. Calculate the phase consistency descriptor of the heterogeneous image, which is used to measure the phase of a specific frequency component in the heterogeneous image; for each pixel, the phase can be calculated by the following formula: ; in, are pixel coordinates, w are different scales, is the phase at scale w, M represents the number of sizes; S2032. Calculate the gradient descriptor of the heterogeneous image; The gradient direction descriptor of the heterogeneous image is as follows: ; ; in, and are the gradients in the x and y directions when the scale is m, respectively, and 𝐼 is the brightness of the image; thus, the gradient descriptor of the heterogeneous image is expressed as:

[0014] ; In the formula, Indicates the direction of the gradient, that is, the direction of the brightness change of the image at the point (x, y); S2033. Set the statistical scale to m, and take the pixel (x, y) as the center, and the gradient histogram of the surrounding 10×10 pixels is expressed as follows: ; S2034. Taking the image target size as the benchmark, take M / 2 scales upward and downward respectively, for a total of M scales; the phase consistency descriptor at each scale , gradient descriptor , gradient histogram It forms a descriptor vector at M scales; S2035. Use the weighted correlation distance method for matching; the expression is as follows: ; Where WCD represents the weighted correlation distance, which is the weighted sum of the differences between the descriptor vectors of the visible light image and the infrared image from scale 1 to scale M; m is the scale of the descriptor; M represents the number of sizes; and Represent the multi-scale descriptor vector at the mth scale in the visible light image and infrared image respectively; is the weight of the mth scale, expressed as: ; is the descriptor vector of the mth scale and The variance of When the WCD value is greater than 0.5, it is considered and The match is successful; the projection relationship between the coordinates of each pixel in the visible light image and the infrared image is stored as matrix A.

[0015] Preferably, the method for adjusting the field of view angle of a single pixel of the visible light image to be equal to the field of view angle of a single pixel of the infrared image in step S202 is as follows: When the UAV performs a mission, if the visible lens is used as the main lens and the infrared is integrated for tracking, the following steps are performed: when the visible lens is used as the main lens and the ground station sends a large field of view or small field of view command, only the visible lens responds; after the ground station control ends, the visible lens is in a stationary state. At this time, the focal length value g of the visible light is read out. CCD ; According to the focal length of visible light g CCD , visible pixel size and infrared pixel size, calculate the infrared focal length of a single pixel under the same field of view, and the camera control board controls the repeated movement of the infrared camera through the PID algorithm until the visible light focal length is 1 / 6 of the infrared focal length; When the UAV is performing a mission, if the infrared lens is the main lens and the visible lens is integrated and tracked, the following steps are performed: when the infrared lens is the main lens and the ground station sends a large field of view or small field of view command, only the infrared lens responds; after the ground station control ends, the infrared lens is in a static state. At this time, the focal length value g of the infrared lens is read out. IR ; According to the focal length value g of the infrared lens IR , according to the visible pixel size and the infrared pixel size, the visible light focal length of a single pixel under the same field of view is calculated. The camera control board controls the repeated movement of the visible light camera through the PID algorithm until the visible light focal length is 1 / 6 of the infrared focal length.

[0016] Preferably, the fusion tracking method in step S3 specifically includes the following sub-steps: S301. Calculate the local contrast of the visible light image and the infrared image; normalize the local contrast of the visible light image and the infrared image to obtain the weight of the local contrast of the visible light image and the infrared image; S302. Calculate the information entropy of the visible light image and the information entropy of the infrared image, normalize the information entropy of the visible light image and the infrared image, and obtain the weights of the information entropy of the visible light image and the infrared image; S303. Convert the RGB format file of the visible light image to YUV, and make the Y channel brightness value of the visible light image Y CCD and infrared image Y channel brightness value Y IR superposition, pixel-by-pixel fusion tracking, and generation of fusion tracking images Y fused , the expression is as follows: ; In the formula, Y fused represents the fused tracking image, Y CCD represents the Y channel brightness value of the visible light image.Y ir is the Y channel brightness value of the infrared image; represents the visible light image weight based on local contrast, represents the infrared image weight based on local contrast; represents the weight of the visible light image based on information entropy, Represents the infrared image weight based on information entropy.

[0017] Preferably, step S301 specifically includes the following sub-steps: S3011. Determine a local area centered on the feature, apply a mean filter to the local area to obtain the local average brightness, calculate the standard deviation of the deviation between the pixel value in the local area and the local average brightness, and take the maximum value of the standard deviation as the local contrast; The calculation formulas for the local standard deviation of visible light images and infrared images are as follows: ; ; In the formula, is the visible image at coordinates ( x , y ), is the infrared image at coordinates ( x , y ), is the local area Ω in the visible light image or infrared image f The average pixel value, |Ω f ∣ represents the local area Ω in the visible light image or infrared image f The total number of pixels; The local contrast expression of a visible light image or infrared image is as follows: CCDLocalContrast=max( ); IRLocalContrast=max( ); S3012. Normalize the local contrast of the visible light image and the infrared image so that the sum of CCDLocalContrast and IRLocalContrast is 1; obtain the weight of the local contrast of the visible light image and the infrared image: ; In the formula, represents the visible light image weight based on local contrast, represents the infrared image weight based on local contrast; The step S302 specifically includes the following sub-steps: S3021. Use a histogram to represent the number of times the gray value i appears in the visible light image or infrared image; for each pixel in the gray image, the gray value or , do the following: ; In the formula, and Represents visible histogram and infrared histogram; i represents grayscale value; represents the indicator function; when ;when ; Represents the grayscale value of the pixel with coordinates (x, y) in the visible image. Represents the grayscale value of the pixel with coordinates (x, y) in the infrared image; S3022. Perform normalization to convert the histogram into a probability distribution by dividing the frequency of each gray value by the total number of pixels in the image: ; ; In the formula, and are the probability distribution of each gray value of the visible light image and infrared image respectively, and are the width and height of the visible light image, respectively. and are the width and height of the infrared image respectively; S3023. Using the Shannon information entropy formula, multiply the probability of each gray value by its logarithm with base 2, and accumulate all 256 possible gray values ​​to calculate the information entropy of the visible light image. H CCD and infrared image information entropy H ir; ; ; S3024. Normalize the information entropy of the visible light image and the infrared image so that H CCD and H ir is 1, and the weights of the information entropy of the visible light image and the infrared image are obtained: ; In the formula, represents the weight of the visible light image based on information entropy, Represents the infrared image weight based on information entropy.

[0018] Preferably, the correlation tracking algorithm in step S3 specifically includes the following steps: S304. Target region initialization: in the fused tracking image, the initial region of the target is determined by the target detection algorithm, and the initial region is used as the starting point of the tracking algorithm; S305. Template matching and correlation calculation: Use the initial area of ​​the target as a template and search for the target by calculating the correlation between the template and each possible position in the image; the correlation can be calculated by the following formula: ; in, R ( x , y ) is at the fused image position ( x , y ), T ( i , j ) is the pixel value of the fused template image, I ( x + i , y + j ) is the pixel value of the target fused image, μT and μI are the means of the template and target regions, respectively; S306. Peak detection and target positioning: In the correlation graph, the new position of the target is determined by finding the local maximum; if the peak value is higher than the preset threshold, the target is considered to be successfully tracked at this position, and the target model is updated according to the tracking result.

[0019] Preferably, the method for performing target search area offset according to ground speed compensation parameters in step S3 specifically comprises the following steps: S307. Compensate value based on ground speed C Calculate the ground speed of the drone in Δt time C Distance moved Δ P offset, the expression is:

[0020] S308. Determine the initial search area center as the target position currently observed by the drone P current ; Set the search area according to Δ P offset Move, the new search area center is: ; S309. Execute the correlation tracking algorithm in the new search area to search for the target with the maximum probability and calculate the target miss distance Δ P error .

[0021] Preferably, step S4 specifically includes the following sub-steps: S401. The servo system contains two closed control loops: speed loop and tracking loop; the speed loop receives the ground speed compensation value C , forming a speed loop including ground speed compensation value; ; Where: Δ P speed_error represents the angular rate error of the speed loop, that is, the difference between the expected angular rate and the actual angular rate; C is the ground speed compensation value; Kp_ speed is the proportional gain of the speed loop; Ki_ speed is the integral gain of the speed loop; Δ P speed_correction is the speed loop control output, which is used to complete the speed loop closed loop; S402. The error input of the tracking loop of the servo system is the miss distance Δ P error ; According to the off-target amount Δ P error Adjust the pointing direction of the optoelectronic turret to reduce deviation and achieve stable tracking: ; In the formula, Δ P correction is the tracking loop control output, which is used to complete the tracking loop closure to achieve accurate tracking of the target. P error represents the off-target amount, Kp , Ki and Kd They are the proportional, integral and differential gains of the tracking loop of the servo control system.

[0022] Compared with the prior art, the present invention can achieve the following beneficial effects: The introduction of the ground speed compensation mechanism significantly improves the tracking stability and accuracy of the UAV on the target in a dynamic flight environment.

[0023] The image registration algorithm based on multi-scale descriptors and optimized search strategy realizes real-time registration and fusion of images and improves the real-time performance of target tracking.

[0024] The fusion tracking method combines local contrast and information entropy, dynamically adjusts the image fusion ratio, maximizes the information content of the fused image, and enhances the accuracy and robustness of target tracking.

[0025] It is suitable for image registration under different resolutions and zoom conditions, which reduces the workload of registration and calibration and improves the efficiency of image registration.

[0026] In summary, the present invention can realize real-time registration and dynamic fusion of visible infrared images. The image features after fusion are rich, and the target features can be effectively extracted. The target search area is offset according to the ground speed compensation parameter, which effectively improves the reliability of the target tracking algorithm. The servo system inputs the ground speed compensation parameter into the speed closed-loop control loop, and combined with the miss amount, the target can be stably tracked, which effectively improves the tracking performance of the UAV for ground targets in a dynamic flight environment. It is suitable for the registration of visible infrared images with different resolutions and under the condition that the visible infrared camera is a zoom lens, and is particularly suitable for the registration of images under the variable focal length visible light camera and the variable focal length infrared camera on the optoelectronic turret of the UAV. It has broad application prospects and practical application value, and significantly reduces the workload of registration and calibration. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 The present invention is a flowchart of a method for stable target tracking combined with ground speed compensation provided in accordance with an embodiment of the present invention.

[0028] Figure 2 It is a visible light image taken by a drone according to an embodiment of the present invention.

[0029] Figure 3 This is an infrared image taken by a drone according to an embodiment of the present invention.

[0030] Figure 4 It is a fused tracking image generated by fusing the visible light image and the infrared image on the drone provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0031] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the following description, the same modules are represented by the same reference numerals. In the case of the same reference numerals, their names and functions are also the same. Therefore, the detailed description thereof will not be repeated.

[0032] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not constitute a limitation of the present invention.

[0033] See also Figure 1 The present invention provides a method for stable target tracking combined with ground speed compensation, which specifically comprises the following steps: S1. Measure the flight speed and flight direction of the drone through the drone-mounted IMU to determine the ground speed; use the angle measurement system of the drone-mounted optoelectronic turret to obtain the angle parameters on the drone's flight path, including the azimuth angle α and the pitch angle β; calculate the ground speed compensation value based on the ground speed and angle parameters; Ground speed compensation refers to compensating the motion of the drone relative to the ground, that is, converting the celestial, easterly and northerly speeds into projections on the ground plane. The calculation method of the ground speed compensation value includes: The drone has an astronomical speed , Eastward speed and northbound speed , and α is the angle between the azimuth of the optoelectronic turret and the flight path of the UAV, β is the pitch angle of the optoelectronic turret, and the ground speed compensation value is calculated based on the above speed components and angles; S101. Increase the eastward speed and northbound speed Converted to a velocity component relative to the drone's flight path (defined by the azimuth angle α) : ; In the formula, α represents the azimuth angle of the UAV flight path, represents the eastward speed, represents the north speed; S102. Adjust the velocity component according to the pitch angle β on the UAV flight path , to take into account the tilt of the drone relative to the horizontal plane, and obtain the adjusted horizontal velocity component, the calculation formula is as follows: ; In the formula, Based on the eastward speed of the drone and northbound speed , the horizontal velocity component calculated from the azimuth angle α and the elevation angle β of the optoelectronic turret; represents the velocity component; β represents the pitch angle on the flight path of the UAV; S103. Calculate the ground speed compensation value; the calculation formula is as follows: ; In the formula, Indicates the ground speed compensation value; is a proportionality constant used to adjust the compensation intensity; Represents the horizontal velocity component.

[0034] Step S1 also includes: inputting the ground speed compensation value into the relevant tracking algorithm to predict the dynamic position change of the target.

[0035] S2. Install the visible light camera and infrared camera on the optoelectronic turret on the drone, and use the optoelectronic turret on the drone to capture visible light images and infrared images (see Figure 2 and Figure 3 ), real-time registration and fusion are performed through image registration algorithm; specifically, the following sub-steps are included: S201. Ground calibration: calibrate the visible light Y-direction field angle of the visible light camera at each focal length and the infrared Y-direction field angle of the infrared camera at each focal length on the ground, record the relationship between the field angle and the focal length as a file, and store it on the camera control board; The measurement and calibration method specifically includes: using a reentrant visible infrared dual light common field light tube (the light tube contains crosshairs, and the step length is 0.1m) to gradually measure the Y-direction field angle of the visible light camera and the Y-direction field angle of the infrared camera; Specifically, the optoelectronic turret is equipped with a measured azimuth and pitch encoder; assuming that the focal length of the visible light camera lens is CCD_F at this time, the pitch angle of the optoelectronic turret is locked to 0, align the upper edge of the visible light camera with the crosshairs of the light tube, and read the azimuth angle CCDup of the optoelectronic turret at this time; rotate the optoelectronic turret to align the lower edge of the visible light camera with the crosshairs of the light tube, and read the azimuth angle CCDdown of the optoelectronic turret at this time, then when the focal length of the visible light camera is CCD_F, the visible light Y-direction field of view angle is CCD_Y=CCDup-CCDdown; record the corresponding relationship between the focal length of the visible light lens and the field of view angle in array c; similarly, measure the Y-direction field of view angle of the infrared camera, IR_Y=IRup-IRdown; record the corresponding relationship between the focal length of the infrared lens and the field of view angle in array d, and store it in the optoelectronic turret recorder on the drone.

[0036] S202. Coarse matching of field of view angle: The camera control board controls the movement of the visible light camera or the infrared camera through the PID algorithm, so that the field of view angle of a single pixel of the visible light image is equal to the field of view angle of a single pixel of the infrared image; the specific operation is as follows: If each pixel of the visible light image and the infrared image has the same field of view, the relationship between the focal lengths of the visible light lens and the infrared lens is derived as follows: Visible light pixel size: ; Infrared pixel size: ; The focal length of the visible light lens is fvi ; The focal length of the infrared lens is fir ; The angular resolution θ can be expressed by the following formula: θ = pixel size / focal length; The angular resolution of the visible light image is θ vis = dvis / fvis ; The angular resolution of the infrared image is ; To make θ vis Equal to θ ir , the following equation can be established: ; That is, when the visible light pixel size is 2.5 μm and the infrared pixel size is 15 μm, that is, the visible focal length is 1 / 6 of the infrared focal length, the field of view of a single pixel in the visible light image is equal to the field of view of a single pixel in the infrared image.

[0037] The visible light camera and infrared camera use external triggering to ensure that the visible and infrared are exposed at the same time.

[0038] The method of adjusting the field of view angle of a single pixel of a visible light image to be equal to the field of view angle of a single pixel of an infrared image, that is, adjusting the visible focal length to 1 / 6 of the infrared focal length, is as follows: When the UAV performs a mission, if the visible lens is used as the main lens and the infrared is integrated for tracking, the following steps are performed: when the visible lens is used as the main lens and the ground station sends a large field of view or small field of view command, only the visible lens responds; after the ground station control ends, the visible lens is in a stationary state. At this time, the focal length value g of the visible light is read out. CCD ; According to the focal length of visible light g CCD , visible pixel size and infrared pixel size, calculate the infrared focal length of a single pixel under the same field of view, and the camera control board controls the repeated movement of the infrared camera through the PID algorithm until the visible light focal length is 1 / 6 of the infrared focal length; When the UAV is performing a mission, if the infrared lens is the main lens and the visible lens is integrated and tracked, the following steps are performed: when the infrared lens is the main lens and the ground station sends a large field of view or small field of view command, only the infrared lens responds; after the ground station control ends, the infrared lens is in a static state. At this time, the focal length value g of the infrared lens is read out. IR ; According to the focal length value g of the infrared lens IR , according to the visible pixel size and the infrared pixel size, the visible light focal length of a single pixel under the same field of view is calculated. The camera control board controls the repeated movement of the visible light camera through the PID algorithm until the visible light focal length is 1 / 6 of the infrared focal length.

[0039] S203. Image descriptor extraction and registration: extract heterogeneous image descriptors of visible light image and infrared image, match the descriptors, store the projection relationship of each pixel coordinate of visible light image and infrared image as matrix A; call matrix A to complete registration; The heterogeneous image descriptors of visible light images and infrared images include phase consistency descriptors and gradient descriptors. The method for matching the descriptors is to extract the common feature descriptors of visible light images and infrared images by combining the phase consistency descriptors and gradient descriptors, and then use the weighted correlation distance (WCD) method for matching. The specific method is as follows: S2031. Calculate the phase consistency descriptor of the heterogeneous image, which is used to measure the phase of a specific frequency component in the heterogeneous image; for each pixel, the phase can be calculated by the following formula: ; in, are pixel coordinates, w are different scales, is the phase at scale w, M represents the number of sizes (usually 10); S2032. Calculate the gradient descriptor of the heterogeneous image; The gradient direction descriptor of the heterogeneous image is as follows: ; ; in, and are the gradients in the x and y directions when the scale is m, respectively, and 𝐼 is the brightness of the image; thus, the gradient descriptor of the heterogeneous image is expressed as:

[0040] ; In the formula, Indicates the direction of the gradient, that is, the direction of the brightness change of the image at the point (x, y); S2033. Set the statistical scale to m, and take the pixel (x, y) as the center, and the gradient histogram of the surrounding 10×10 pixels is expressed as follows: ; S2034. Based on the image target size, take M / 2 scales upward and M / 2 scales downward, for a total of M scales. Phase consistency descriptor under , gradient descriptor (gradient magnitude) , gradient histogram It constitutes the descriptor vector at the mth scale : ; The visible light image and infrared image are operated according to the above method, and each feature point in the obtained visible light image and infrared image generates a rich multi-scale descriptor vector and ,This descriptor effectively describes the local structural information of visible and infrared images, and provides a solid foundation for feature matching.

[0041] S2035. Use the weighted correlation distance (WCD) method for matching; the expression is as follows: ; Where WCD represents the weighted correlation distance, which is the weighted sum of the differences between the descriptor vectors of the visible light image and the infrared image from scale 1 to scale M; m is the scale of the descriptor (the value of m ranges from 1 to M); M represents the number of sizes (usually 10); and Represent the multi-scale descriptor vector at the mth scale in the visible light image and infrared image respectively; is the weight of the mth scale, expressed as: ; is the descriptor vector of the mth scale and The variance of When the WCD value is greater than 0.5, it is considered and The match is successful; the projection relationship between the coordinates of each pixel in the visible light image and the infrared image is stored as matrix A.

[0042] Principle: After step S202, the visible and infrared fields of view are basically consistent in the Y direction in theory. However, due to the field of view angle calibration error in step S1 and the focus control error in step S202, there are still slight differences in the visible and infrared fields of view, so registration is required. The calculation of the common feature descriptors of visible light images and infrared images is a key step in the matching of visible light images and infrared images. It provides a unique vector representation for each feature point of the visible light image and the infrared image.

[0043] The weighted correlation distance (WCD) method not only considers the Euclidean distance between feature vectors, but also the correlation between feature dimensions. This method is particularly suitable for situations where there are intrinsic connections between feature dimensions.

[0044] Because visible infrared uses external triggering to ensure that visible infrared is exposed at the same time, after exposure, a new frame of visible image and infrared image arrive at the onboard computer. Only by calling matrix A can the real-time registration of visible infrared be completed, so that the real-time performance of the registration is effectively guaranteed.

[0045] S3. Extract local contrast and information entropy based on the registered fused image, and dynamically adjust the fusion ratio of the visible light image and the infrared image through the fusion tracking method to generate a fused tracking image ( Figure 4); Input the fused tracking image into the relevant tracking algorithm to track the target; offset the target search area according to the ground speed compensation parameters to ensure that the target is searched with the maximum probability; The fusion tracking method specifically includes: S301. Calculate the local contrast of the visible light image and the infrared image; normalize the local contrast of the visible light image and the infrared image to obtain the weight of the local contrast of the visible light image and the infrared image; specifically include the following sub-steps: S3011. Determine a local area centered on the feature, apply a mean filter to the local area to obtain the local average brightness, calculate the standard deviation of the deviation between the pixel values ​​in the local area and the local average brightness, and take the maximum value of the standard deviation as the local contrast.

[0046] Local standard deviation It is a local area The measure of the internal pixel value fluctuation, the local standard deviation calculation formulas for visible light images and infrared images are as follows: ; ; In the formula, is the visible image at coordinates ( x , y ), is the infrared image at coordinates ( x , y ), is the local area Ω in the visible light image or infrared image f The average pixel value, |Ω f ∣ represents the local area Ω in the visible light image or infrared image f The total number of pixels.

[0047] The local standard deviation is calculated by calculating the local region Ω f Inner pixel value and its average value μf The sum of the squared deviations of , and then take the square root. This value reflects the dispersion of pixel intensity in the feature area and is an indicator of the complexity of local texture.

[0048] The local contrast of a visible light image or infrared image is defined as the maximum value of its local standard deviation: CCDLocalContrast=max( ); IRLocalContrast=max( ); S3012. Normalize the local contrast of the visible light image and the infrared image so that the sum of CCDLocalContrast and IRLocalContrast is 1; obtain the weight of the local contrast of the visible light image and the infrared image: ; In the formula, represents the visible light image weight based on local contrast, Represents the infrared image weight based on local contrast.

[0049] S302. Calculate the information entropy of the visible light image and the information entropy of the infrared image, normalize the information entropy of the visible light image and the infrared image, and obtain the weights of the information entropy of the visible light image and the infrared image; specifically, the following sub-steps are included: S3021. Use a histogram to represent the number of times the gray value i appears in the visible light image or infrared image; for each pixel in the gray image, the gray value or , do the following: ; In the formula, and Represents visible histogram and infrared histogram; i represents grayscale value; represents the indicator function; when ;when ; Represents the grayscale value of the pixel with coordinates (x, y) in the visible image. Represents the grayscale value of the pixel with coordinates (x, y) in the infrared image; S3022. Perform normalization to convert the histogram into a probability distribution by dividing the frequency of each gray value by the total number of pixels in the image: ; ; In the formula, and are the probability distribution of each gray value of the visible light image and infrared image respectively, and are the width and height of the visible light image, respectively. and are the width and height of the infrared image respectively.

[0050] S3023. Using the Shannon information entropy formula, multiply the probability of each gray value by its logarithm with base 2, and accumulate all 256 possible gray values ​​to calculate the information entropy of the visible light image. H CCD and infrared image information entropyH ir; ; ; The range of information entropy is from 0 to a maximum of 8 bits. This range reflects the amount of information in an image from complete disorder (the grayscale value of each pixel is random and has equal probability) to complete order (all pixels have the same grayscale value); when all pixels have the same grayscale value (that is, the image is completely uniform), H CCD = 0, H ir=0; when the probability of each gray value appearing is equal, the information entropy H CCD = 8, H ir=8, reaching the maximum value.

[0051] S3024. Normalize the information entropy of the visible light image and the infrared image so that H CCD and H ir is 1, and the weights of the information entropy of the visible light image and the infrared image are obtained: ; In the formula, represents the weight of the visible light image based on information entropy, Represents the infrared image weight based on information entropy.

[0052] S303. Convert the RGB format file of the visible light image to YUV, and make the Y channel brightness value of the visible light image Y CCD and infrared image Y channel brightness value Y IR superposition, pixel-by-pixel fusion tracking, and generation of fusion tracking images Y fused , the expression is as follows: ; In the formula, Y fused represents the fused tracking image, Y CCD represents the Y channel brightness value of the visible light image. Y ir is the Y channel brightness value of the infrared image; represents the visible light image weight based on local contrast, represents the infrared image weight based on local contrast; represents the weight of the visible light image based on information entropy, Represents the infrared image weight based on information entropy.

[0053] In step S3, the relevant tracking algorithm specifically includes the following steps: S304. Target region initialization: in the fused tracking image, the initial region of the target is determined by the target detection algorithm, and the initial region is used as the starting point of the tracking algorithm; S305. Template matching and correlation calculation: Use the initial area of ​​the target as a template and search for the target by calculating the correlation between the template and each possible position in the image; the correlation can be calculated by the following formula: ; in, R ( x , y ) is at the fused image position ( x , y ), T ( i , j ) is the pixel value of the fused template image, I ( x + i , y + j ) is the pixel value of the target fused image, μT and μI are the means of the template and target regions, respectively.

[0054] Due to the limited computing power of the embedded airborne computer, it is impossible to search the entire area of ​​the entire frame. The traditional method is to search a limited area near the location where the target appeared in the previous frame.

[0055] S306. Peak detection and target positioning: In the correlation graph, the new position of the target is determined by finding the local maximum; if the peak is higher than the preset threshold, the target is considered to be successfully tracked at this position, and the target model is updated according to the tracking results, including the shape, size and appearance characteristics of the target.

[0056] In step S3, the method for performing target search area offset according to the ground speed compensation parameters specifically includes the following steps: S307. Compensate value based on ground speed C Calculate the ground speed of the drone in Δt time C Distance moved Δ P offset, the expression is: ; S308. Determine the initial search area center as the target position currently observed by the drone P current ; Set the search area according to Δ P offset Move, the new search area center is: ; S309. Execute the correlation tracking algorithm in the new search area to search for the target with the maximum probability and calculate the target miss distance Δ P error .

[0057] S4. The servo system inputs the ground speed compensation parameter into the speed closed-loop control loop, and combines the miss distance of the relevant tracking to stably track the target; specifically, the following sub-steps are included: S401. The servo system contains two closed control loops: speed loop and tracking loop; the speed loop receives the ground speed compensation value C , forming a speed loop including ground speed compensation value; ; Where: Δ P speed_error represents the angular rate error of the speed loop, that is, the difference between the expected angular rate and the actual angular rate. This difference is measured by the angular rate gyro built into the optoelectronic turret. The angular rate gyro can only measure the change in the angular velocity of the optoelectronic turret, but cannot measure the linear velocity change caused by the flight of the drone. C is the ground speed compensation value, that is, the linear velocity change caused by the flight of the drone. Δ P speed_error+ C It represents the angular velocity of the optoelectronic turret + the linear velocity generated by the UAV flight, which can fully reflect the speed loop error of the optoelectronic turret in actual flight; by adding the ground speed compensation value C to the error term Δ P speed_error, you can ensure that the speed loop takes into account the linear speed of the drone when calculating the correction, thereby improving the accuracy of the speed loop's responsiveness; Kp_ speed is the proportional gain of the speed loop; Ki_ speed is the integral gain of the speed loop; Δ P speed_correction is the speed loop control output, which is used to complete the speed loop closed loop.

[0058] S402. The error input of the tracking loop of the servo system is the miss distance Δ P error ; According to the off-target amount Δ P error Adjust the pointing direction of the optoelectronic turret to reduce deviation and achieve stable tracking: ; In the formula, Δ P correction is the tracking loop control output, which is used to complete the tracking loop closure to achieve accurate tracking of the target. P error represents the off-target amount, Kp , Ki and KdThey are the proportional, integral and differential gains of the tracking loop of the servo control system.

[0059] The core of the present invention is to introduce a ground speed compensation mechanism, which calculates the ground speed of the drone in real time and uses it as an input parameter of the tracking algorithm to predict the dynamic changes of the target and achieve stable tracking of the target. First, the flight speed and direction of the drone are obtained through the navigation system of the drone, and then the ground speed vector is calculated based on the azimuth and pitch angle of the airborne optoelectronic turret; the ground speed vector is then used to adjust the relevant tracking algorithm to ensure that the target tracking system can accurately predict the changes in the target position during the flight of the drone.

[0060] The present invention also includes an image registration algorithm based on multi-scale descriptors and optimized search strategies, which can run in real time on an onboard computer to further improve the accuracy of target tracking. In addition, a fusion tracking method is provided, which combines local contrast and information entropy to dynamically adjust the fusion ratio of visible light images and infrared images to maximize the information content of the fused image.

[0061] The advantage of the present invention is that it is not only applicable to image registration under different resolutions and zoom conditions, but also significantly reduces the workload of registration calibration and improves the efficiency of image registration. By integrating tracking technology, the accuracy and robustness of target tracking are enhanced, meeting the needs of real-time image processing of drones, and has broad application prospects and practical application value.

[0062] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the disclosure of the present invention can be performed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and this document does not limit this.

[0063] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for stable target tracking combined with ground speed compensation, characterized in that: The specific steps include: S1. Measure the flight speed and direction of the drone through the drone-mounted IMU to determine the ground speed; use the angle measurement system of the drone-mounted optoelectronic turret to obtain the angle parameters on the drone's flight path; calculate the ground speed compensation value based on the ground speed and angle parameters; S2. Install both the visible light camera and the infrared camera on the optoelectronic turret on the drone, use the optoelectronic turret on the drone to capture visible light images and infrared images, and perform real-time registration and fusion through image registration algorithm; S3. Extract local contrast and information entropy based on the registered fused image, and dynamically adjust the fusion ratio of the visible light image and the infrared image through the fusion tracking method to generate a fused tracking image; input the fused tracking image into the relevant tracking algorithm to track the target; offset the target search area based on the parameters of the ground speed compensation to ensure that the target is searched with the maximum probability; S4. The servo system inputs the ground speed compensation parameter into the speed closed-loop control loop, and combines it with the miss distance of the relevant tracking to stably track the target.

2. The method for stable target tracking combined with ground speed compensation according to claim 1, characterized in that: The angle parameters in step S1 include an azimuth angle α and a pitch angle β; The method for calculating the ground speed compensation value specifically includes: S101. Define the eastward speed of the drone by the azimuth angle α and northbound speed ; Set the drone's eastward speed and northbound speed Converted to velocity component relative to the UAV flight path : ； In the formula, α represents the azimuth angle of the UAV flight path, represents the eastward speed, represents the north speed; S102. Adjust the velocity component according to the pitch angle β on the UAV flight path , and the adjusted horizontal velocity component is obtained. The calculation formula is as follows: ； In the formula, Based on the eastward speed of the drone and northbound speed , the horizontal velocity component calculated from the azimuth angle α and the pitch angle β; represents the velocity component; β represents the pitch angle on the flight path of the UAV; S103. Calculate the ground speed compensation value; the calculation formula is as follows: ； In the formula, Indicates the ground speed compensation value; is a proportionality constant used to adjust the compensation intensity; Represents the horizontal velocity component.

3. The method for stable target tracking combined with ground speed compensation according to claim 1, characterized in that: The step S2 specifically includes the following sub-steps: S201. Ground calibration: calibrate the visible light Y-direction field angle of the visible light camera at each focal length and the infrared Y-direction field angle of the infrared camera at each focal length on the ground, record the relationship between the field angle and the focal length as a file, and store it on the camera control board; S202. Coarse matching of field of view angle: The camera control board controls the movement of the visible light camera or the infrared camera through the PID algorithm so that the field of view angle of a single pixel of the visible light image is equal to the field of view angle of a single pixel of the infrared image; S203. Image descriptor extraction and registration: extract heterogeneous image descriptors of the visible light image and the infrared image, match the descriptors, and store the projection relationship of each pixel coordinate of the visible light image and the infrared image as a matrix A; Call matrix A to complete the alignment.

4. The method for stable target tracking combined with ground speed compensation according to claim 3, characterized in that: In step S203, the heterogeneous image descriptors of the visible light image and the infrared image include a phase consistency descriptor and a gradient descriptor; the weighted correlation distance method is used for matching, and the specific method is as follows: S2031. Calculate the phase consistency descriptor of the heterogeneous image, which is used to measure the phase of a specific frequency component in the heterogeneous image; for each pixel, the phase can be calculated by the following formula: ； in, are pixel coordinates, w are different scales, is the phase at scale w, M represents the number of sizes; S2032. Calculate the gradient descriptor of the heterogeneous image; The gradient direction descriptor of the heterogeneous image is as follows: ； ； in, and are the gradients in the x and y directions when the scale is m, respectively, and 𝐼 is the brightness of the image; thus, the gradient descriptor of the heterogeneous image is expressed as: ； In the formula, Indicates the direction of the gradient, that is, the direction of the brightness change of the image at the point (x, y); S2033. Set the statistical scale to m, and take the pixel (x, y) as the center, and the gradient histogram of the surrounding 10×10 pixels is expressed as follows: ； S2034. Taking the image target size as the benchmark, take M / 2 scales upward and downward respectively, for a total of M scales; the phase consistency descriptor at each scale , gradient descriptor , gradient histogram It forms a descriptor vector at M scales; S2035. Use the weighted correlation distance method for matching; the expression is as follows: ； Where WCD represents the weighted correlation distance, which is the weighted sum of the differences between the descriptor vectors of the visible light image and the infrared image from scale 1 to scale M; m is the scale of the descriptor; M represents the number of sizes; and Represent the multi-scale descriptor vector at the mth scale in the visible light image and infrared image respectively; is the weight of the mth scale, expressed as: ; is the descriptor vector of the mth scale and The variance of When the WCD value is greater than 0.5, it is considered and The match is successful; the projection relationship between the coordinates of each pixel in the visible light image and the infrared image is stored as matrix A.

5. The method for stable target tracking combined with ground speed compensation according to claim 3, characterized in that: The method for adjusting the field of view angle of a single pixel of the visible light image to be equal to the field of view angle of a single pixel of the infrared image in step S202 is as follows: When the UAV performs a mission, if the visible lens is used as the main lens and the infrared is integrated for tracking, the following steps are performed: when the visible lens is used as the main lens and the ground station sends a large field of view or small field of view command, only the visible lens responds; after the ground station control ends, the visible lens is in a stationary state. At this time, the focal length value g of the visible light is read out. CCD ; According to the focal length of visible light g CCD , visible pixel size and infrared pixel size, calculate the infrared focal length of a single pixel under the same field of view, and the camera control board controls the repeated movement of the infrared camera through the PID algorithm until the visible light focal length is 1 / 6 of the infrared focal length; When the UAV is performing a mission, if the infrared lens is the main lens and the visible lens is integrated and tracked, the following steps are performed: when the infrared lens is the main lens and the ground station sends a large field of view or small field of view command, only the infrared lens responds; after the ground station control ends, the infrared lens is in a static state. At this time, the focal length value g of the infrared lens is read out. IR ; According to the focal length value g of the infrared lens IR , according to the visible pixel size and the infrared pixel size, the visible light focal length of a single pixel under the same field of view is calculated. The camera control board controls the repeated movement of the visible light camera through the PID algorithm until the visible light focal length is 1 / 6 of the infrared focal length.

6. The method for stable target tracking combined with ground speed compensation according to claim 1, characterized in that: The fusion tracking method in step S3 specifically includes the following sub-steps: S301. Calculate the local contrast of the visible light image and the infrared image; normalize the local contrast of the visible light image and the infrared image to obtain the weight of the local contrast of the visible light image and the infrared image; S302. Calculate the information entropy of the visible light image and the information entropy of the infrared image, normalize the information entropy of the visible light image and the infrared image, and obtain the weights of the information entropy of the visible light image and the infrared image; S303. Convert the RGB format file of the visible light image to YUV, and make the Y channel brightness value of the visible light image Y CCD and infrared image Y channel brightness value Y IR superposition, pixel-by-pixel fusion tracking, and generation of fusion tracking images Y fused , the expression is as follows: ； In the formula, Y fused represents the fused tracking image, Y CCD represents the Y channel brightness value of the visible light image. Y ir is the Y channel brightness value of the infrared image; represents the visible light image weight based on local contrast, represents the infrared image weight based on local contrast; represents the weight of the visible light image based on information entropy, Represents the infrared image weight based on information entropy.

7. The method for stable target tracking combined with ground speed compensation according to claim 6, characterized in that: The step S301 specifically includes the following sub-steps: S3011. Determine a local area centered on the feature, apply a mean filter to the local area to obtain the local average brightness, calculate the standard deviation of the deviation between the pixel value in the local area and the local average brightness, and take the maximum value of the standard deviation as the local contrast; The calculation formulas for the local standard deviation of visible light images and infrared images are as follows: ； ； In the formula, is the visible image at coordinates ( x , y ), is the infrared image at coordinates ( x , y ), is the local area Ω in the visible light image or infrared image f The average pixel value, |Ω f ∣ represents the local area Ω in the visible light image or infrared image f The total number of pixels; The local contrast expression of a visible light image or infrared image is as follows: CCDLocalContrast=max( ); IRLocalContrast=max( ); S3012. Normalize the local contrast of the visible light image and the infrared image so that the sum of CCDLocalContrast and IRLocalContrast is 1; obtain the weight of the local contrast of the visible light image and the infrared image: ； In the formula, represents the visible light image weight based on local contrast, represents the infrared image weight based on local contrast; The step S302 specifically includes the following sub-steps: S3021. Use a histogram to represent the number of times the gray value i appears in the visible light image or infrared image; for each pixel in the gray image, the gray value or , do the following: ； In the formula, and Represents visible histogram and infrared histogram; i represents grayscale value; represents the indicator function; when ;when ; Represents the grayscale value of the pixel with coordinates (x, y) in the visible image. Represents the grayscale value of the pixel with coordinates (x, y) in the infrared image; S3022. Perform normalization to convert the histogram into a probability distribution by dividing the frequency of each gray value by the total number of pixels in the image: ； ； In the formula, and are the probability distribution of each gray value of the visible light image and infrared image, and are the width and height of the visible light image, respectively. and are the width and height of the infrared image respectively; S3023. Using the Shannon information entropy formula, multiply the probability of each gray value by its logarithm with base 2, and accumulate all 256 possible gray values ​​to calculate the information entropy of the visible light image. H CCD and infrared image information entropy H ir; ； ； S3024. Normalize the information entropy of the visible light image and the infrared image so that H CCD and H ir is 1, and the weights of the information entropy of the visible light image and the infrared image are obtained: ； In the formula, represents the weight of the visible light image based on information entropy, Represents the infrared image weight based on information entropy.

8. The method for stable target tracking combined with ground speed compensation according to claim 7, characterized in that: The correlation tracking algorithm in step S3 specifically includes the following steps: S304. Target region initialization: in the fused tracking image, the initial region of the target is determined by the target detection algorithm, and the initial region is used as the starting point of the tracking algorithm; S305. Template matching and correlation calculation: Use the initial area of ​​the target as a template and search for the target by calculating the correlation between the template and each possible position in the image; the correlation can be calculated by the following formula: ； in, R ( x , y ) is at the fused image position ( x , y ), T ( i , j ) is the pixel value of the fused template image, I ( x + i , y + j ) is the pixel value of the target fused image, μT and μI are the means of the template and target regions, respectively; S306. Peak detection and target positioning: In the correlation graph, the new position of the target is determined by finding the local maximum; if the peak value is higher than the preset threshold, the target is considered to be successfully tracked at this position, and the target model is updated according to the tracking result.

9. The method for stable target tracking combined with ground speed compensation according to claim 8, characterized in that: The method for performing target search area offset according to the ground speed compensation parameters in step S3 specifically comprises the following steps: S307. Compensate value based on ground speed C Calculate the ground speed of the drone in Δt time C Distance moved Δ P offset, the expression is: S308. Determine the initial search area center as the target position currently observed by the drone P current ; Set the search area according to Δ P offset Move, the new search area center is: ； S309. Execute the correlation tracking algorithm in the new search area to search for the target with the maximum probability and calculate the target miss distance Δ P error .

10. The method for stable target tracking combined with ground speed compensation according to claim 1, characterized in that: The step S4 specifically includes the following sub-steps: S401. The servo system contains two closed control loops: speed loop and tracking loop; the speed loop receives the ground speed compensation value C , forming a speed loop including ground speed compensation value; ； Where: Δ P speed_error represents the angular rate error of the speed loop, that is, the difference between the expected angular rate and the actual angular rate; C is the ground speed compensation value; Kp_ speed is the proportional gain of the speed loop; Ki_ speed is the integral gain of the speed loop; Δ P speed_correction is the speed loop control output, which is used to complete the speed loop closed loop; S402. The error input of the tracking loop of the servo system is the miss distance Δ P error ; According to the off-target amount Δ P error Adjust the pointing direction of the optoelectronic turret to reduce deviation and achieve stable tracking: ； In the formula, Δ P correction is the tracking loop control output, which is used to complete the tracking loop closure to achieve accurate tracking of the target. P error represents the off-target amount, Kp , Ki and Kd They are the proportional, integral and differential gains of the tracking loop of the servo control system.

Images