A dual mode acquisition tracking method

Abstract

The application discloses a dual-mode capture tracking method and belongs to the technical field of space laser communication. The method can dynamically obtain image information under high exposure and spot information under low exposure through an adaptive exposure and dynamic filter switching mechanism. The technology is suitable for a case where a platform image template of an opposite party can be seen at a short distance. The advantage of the technology is that a wide field of view image can be used for fast target area locking in an initial capture stage, and the technology is favorable for shortening a link establishment time. Meanwhile, when a spot tracking is lost due to interference, the system can be immediately switched to an image mode, and the technology is favorable for fast re-capturing of a target in a local field of view, so that the system can avoid returning to a global blind scanning. The application effectively overcomes the defects of a low capture efficiency, a poor environmental robustness and a slow lock loss recovery of a traditional single-mode APT system, and provides key technical support for a next-generation high-reliability space laser communication.

Description

Technical Field

[0001] This invention belongs to the field of space laser communication technology, specifically relating to a dual-mode acquisition and tracking method, applicable to the acquisition, alignment and tracking (APT) system of laser communication terminals, and particularly suitable for scenarios requiring rapid link establishment and robustness enhancement in dynamic environments. Background Technology

[0002] Acquisition, alignment, and tracking are crucial for establishing and maintaining a stable communication link in space laser communication systems. Traditional APT systems generally employ single-mode laser beacon acquisition and tracking technology. The typical workflow is as follows: the transmitting end's beacon laser beam performs a point-by-point "blind scan" within a pre-defined uncertainty area according to a predetermined scanning trajectory (such as a spiral or grating); the receiving end's detector continuously searches for the light signal, and once a beacon spot is detected, acquisition is considered successful, and the system transitions to a fine-tracking phase to maintain the link.

[0003] However, this acquisition and tracking mode, which relies on a single laser spot beacon, has the following inherent drawbacks, which severely limit the system's performance and reliability in dynamic and complex environments:

[0004] (1) Low capture efficiency and slow link establishment: Traditional systems usually use narrowband filters with low camera exposure to perform a large-scale mechanical "blind scan" of light spots in dark backgrounds. The process is time-consuming and has large errors, making it difficult to quickly detect light spot targets, thus failing to meet the requirements of high-speed communication systems for fast link establishment.

[0005] (2) Poor environmental robustness and weak recovery capability: In complex atmospheric environments (such as turbulence, clouds and fog) or in the presence of instantaneous occlusion and strong background light interference, the single spot mode is prone to attenuation, flickering or loss. Moreover, once tracking is lost due to target maneuvering, strong interference or platform vibration, the system usually needs to return to the global blind scan process, causing a long-term interruption of the communication link, which seriously affects the availability and robustness of the system. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a dual-mode acquisition and tracking method. By fusing image mode and spot mode, it can achieve efficient acquisition and robust tracking of laser communication terminals, and significantly improve link establishment speed and lock-out recovery capability.

[0007] The technical problem addressed by this invention is solved as follows:

[0008] A dual-mode capture and tracking method, based on a visible light / near-infrared camera and an electrically switched filter, includes an image scanning stage, an image capture stage, an image tracking stage, and a spot tracking stage;

[0009] (a) Image scanning stage

[0010] The electrical switching filter is switched to a visible / near-infrared broadband filter, and the visible / near-infrared camera adopts an image adaptive high exposure mode; a rectangular spiral scan is performed in a pre-set area of ​​uncertain angle;

[0011] Real-time detection of the weighted average gray value of the image within the field of view The system determines whether a beacon spot has appeared. If the weighted average gray value exceeds 210, it is determined that a beacon spot has been detected, and the process proceeds to the spot tracking stage.

[0012] (ii) Image capture stage

[0013] During image scanning, a preset target grayscale template is simultaneously identified and matched, and the correlation coefficient between the currently scanned image to be matched and the target grayscale template is calculated.

[0014] When the correlation coefficient is ≤0.8, continue image scanning;

[0015] When the correlation coefficient exceeds 0.8, the image is considered successfully captured; if the weighted average gray value of the currently scanned image exceeds 210, the beacon spot is detected and the process proceeds to the spot tracking stage; if the beacon spot is not detected, the process proceeds to the image tracking stage.

[0016] (III) Image Tracking Stage

[0017] The KCF tracking method is used to continuously track the captured image target; a window is opened at the target location in the image to detect light spots, and the average gray value within the window is calculated. ;like Once the beacon spot is detected, the process transitions to the spot tracking phase.

[0018] (iv) Spot tracking stage

[0019] The electronically switched filter is switched to a near-infrared narrowband filter. The visible light / near-infrared camera adopts a spot-adaptive low-exposure mode, activates the spot mode adaptive exposure adjustment mechanism, and uses the centroid tracking method to extract the spot position.

[0020] Furthermore, during the image scanning phase, the weighted average gray value of the image within the field of view... for:

[0021]

[0022] Where 1≤n≤N, and N is the number of image pixels in the field of view; , Let n be the weight of the nth pixel. Let be the grayscale value of the nth pixel in the image.

[0023] Furthermore, in the image scanning stage, the exposure time adjustment process in the adaptive exposure adjustment mechanism of the visible / near-infrared camera image mode is represented as follows:

[0024]

[0025] in, To adjust the exposure time, This is the current exposure time. As an exposure factor, The weighted average gray value of the image within the current field of view. With setting reference grayscale value The difference between them.

[0026] Furthermore, in the image capture stage, a normalized cross-correlation method is used for the identification and matching of target grayscale templates, and the correlation coefficient is... for:

[0027]

[0028] Where (x, y) is the position of the target grayscale template; T represents the target grayscale template, and I represents the current scanned image to be matched. Both have a width of w and a height of h, 1≤i≤w, 1≤j≤h; The average pixel grayscale value of the scanned image to be matched. The average pixel grayscale value of the target grayscale template.

[0029] Furthermore, in the spot tracking stage, the adaptive exposure adjustment mechanism for the spot mode is specifically as follows:

[0030] Calculate the average gray value of pixels within the window. ;

[0031] An error-based nonlinear step size adjustment strategy is used to adjust the exposure time, expressed as:

[0032]

[0033] in, Indicates the exposure adjustment step size. Represents the average gray value Compared with reference value The difference between them.

[0034] Furthermore, in the spot tracking stage, the spot position is extracted using the centroid tracking method, and the spot centroid ( , )for:

[0035]

[0036] Where 1≤k≤K, 1≤l≤L, and K and L are the width and height of the window in pixels, respectively; Representing pixels The x and y coordinates, Represents pixels The grayscale value; if , then let , The set grayscale threshold; if , then let .

[0037] Furthermore, during the spot tracking stage, if the spot loses its lock due to interference, the electrically switched filter is switched to a visible / near-infrared broadband filter, the visible / near-infrared camera adopts an image adaptive high exposure mode, and the exposure time is increased, returning to the image tracking stage;

[0038] Increase the adjusted exposure time :

[0039] .

[0040] Furthermore, an inverse model feedforward PID control strategy is introduced, and an inverse model feedforward PID control module, including a PID controller and a feedforward controller, is connected in front of the servo turntable.

[0041] The transfer function of the PID controller is:

[0042]

[0043] in, This is the proportionality coefficient. The integral coefficient is... Here are the differential coefficients, and s is the Laplace field operator;

[0044] Feedforward controller The regularized inverse model of the second-order low-pass filter is represented as follows:

[0045]

[0046] in, For the transfer function of the servo turntable, , The filter time constant is The damping ratio is denoted as .

[0047] The beneficial effects of this invention are:

[0048] The method described in this invention utilizes an adaptive exposure and dynamic filter switching mechanism to enable the system to dynamically acquire image information under high exposure and spot information under low exposure. This technology is suitable for situations where the distance is not far and the image template of the other party's platform (such as drones, ground terminals, etc.) can be seen. Its advantages lie in the fact that the system can quickly lock onto the target area using a wide field-of-view image during the initial acquisition phase, which helps shorten the link establishment time. Simultaneously, when spot tracking loses lock due to interference, the system can immediately switch to the image mode, facilitating rapid re-acquisition of the target in a local field of view, thus avoiding a return to global blind scanning. The method described in this invention significantly shortens the laser communication link establishment time through dual-mode deep fusion and intelligent scheduling. Experiments show that the average link establishment time is reduced from 2.8 seconds in traditional methods to 0.7 seconds; the average image acquisition success rate reaches 96.67%, and the overall acquisition probability is improved by approximately 3.97%; after spot tracking loses lock, the re-establishment speed is shortened from 2.8 seconds to 0.5 seconds, and the success rate is improved from 82% to 98%; after introducing inverse model feedforward control, the system's dynamic response is significantly improved, with rise time and settling time reduced by more than 60%. This invention effectively overcomes the shortcomings of traditional single-mode APT systems, such as low acquisition efficiency, poor environmental robustness, and slow recovery from lock-off, providing key technical support for next-generation high-reliability space laser communication. Attached Figure Description

[0049] Figure 1 This is a diagram of the tracking system architecture in the method described in this invention;

[0050] Figure 2 This is a general flowchart of the method described in this invention;

[0051] Figure 3 This is a schematic diagram of the rectangular spiral scanning method described in this invention;

[0052] Figure 4 This is a schematic diagram of the KCF tracking algorithm in the method described in this invention;

[0053] Figure 5 This is a schematic diagram of the centroid tracking algorithm in the method described in this invention;

[0054] Figure 6 This is a block diagram illustrating the implementation of the inverse model feedforward PID control strategy in the method described in this invention. Detailed Implementation

[0055] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0056] This embodiment provides a dual-mode acquisition and tracking method, implemented based on a dual-mode acquisition and tracking unit, including a visible light / near-infrared camera, an electrically switched filter, and a coarse tracking control unit. The symbol / represents OR, and its architecture diagram is shown below. Figure 1As shown, without adding additional hardware, dual-mode capture and tracking is achieved through the camera's adaptive exposure adjustment and dynamic filter switching mechanism, enabling the system to dynamically acquire image information under high exposure and light spot information under low exposure.

[0057] The overall flow of the method described in this embodiment is as follows: Figure 2 As shown, it is divided into four stages: image scanning stage, image capture stage, image tracking stage, and spot tracking stage.

[0058] (a) Image scanning stage:

[0059] Initially, both ends are unlocked, the light spot is invisible, and the image template is not in the field of view. The coarse tracking control unit controls the electrically switched filter to a visible / near-infrared broadband filter, and controls the visible / near-infrared camera to adopt an image adaptive high-exposure mode; a rectangular spiral scan is performed within a pre-set area of ​​angular uncertainty, the scanning method as follows: Figure 3 As shown, quickly search for the target.

[0060] Real-time detection of the average gray value of the image within the field of view The system determines whether a beacon spot has appeared. If the weighted average gray value of the current scanned image exceeds 210, it determines that a beacon spot has been detected and proceeds to the spot tracking stage.

[0061] In this embodiment, in order to adapt to different weather conditions, the adaptive exposure adjustment mechanism of the visible light / near infrared camera image mode is activated simultaneously during scanning. The average gray value is calculated based on the weighted statistical calculation of the overall gray value in the entire field of view. By reducing the weight of extremely dark (possibly background) and overly bright pixels (possibly overexposed areas), background interference is suppressed, making the mean value closer to the effective brightness of the target area, thereby optimizing the exposure parameters.

[0062] The adaptive exposure adjustment mechanism for visible / near-infrared camera image modes is as follows:

[0063] Weighted average gray value of the image across the entire field of view for:

[0064]

[0065] Where 1≤n≤N, and N is the number of image pixels in the field of view; , Let n be the weight of the nth pixel. The image grayscale value of the nth pixel;

[0066] The process of adjusting the exposure time is represented as follows:

[0067]

[0068] in, To adjust the exposure time, This is the current exposure time. As an exposure factor, The weighted average gray value of the image within the current field of view. With setting reference grayscale value The difference between them; in this embodiment, Set to 0.05 to set the reference grayscale value. Set to 100-120.

[0069] (ii) Image capture stage

[0070] During image scanning, a preset target grayscale template is simultaneously identified and matched.

[0071] The normalized cross-correlation method is used for the identification and matching of target grayscale templates, and the correlation coefficient is... for:

[0072]

[0073] Where (x, y) is the position of the target grayscale template; T represents the target grayscale template, and I represents the current scanned image to be matched. Both have a width of w and a height of h, 1≤i≤w, 1≤j≤h; The average pixel grayscale value of the scanned image to be matched. The average pixel grayscale value of the target grayscale template.

[0074] When the correlation coefficient is ≤0.8, continue image scanning;

[0075] When the correlation coefficient exceeds 0.8, the image is considered successfully captured. If the weighted average gray value of the currently scanned image exceeds 210, the beacon spot is detected, and the process proceeds to the spot tracking stage. If the beacon spot is not detected, the process proceeds to the image tracking stage.

[0076] (III) Image Tracking Stage

[0077] The KCF tracking method is used to continuously track the captured image target; the position of the other platform is accurately located, and this information is used to generate an updated and more accurate local template for matching and tracking, so as to achieve stable coarse alignment at both ends.

[0078] A window is opened at the target location in the image to detect the light spot, with the window size set to 50×50 pixels; the average gray value within the window is recorded as follows. ;like Once the beacon spot is detected, the process transitions to the spot tracking phase.

[0079] (iv) Spot tracking stage

[0080] The coarse tracking control unit controls the electrically switched filter to a near-infrared narrowband filter. The coarse tracking control unit controls the visible light / near-infrared camera to adopt a spot adaptive low exposure mode, activates the spot mode adaptive exposure adjustment mechanism, and uses the centroid tracking method to extract the spot position.

[0081] The adaptive exposure adjustment mechanism for the spot mode is as follows:

[0082] Calculate the average gray value of pixels within the window. (All weights are 1);

[0083] An error-based nonlinear step size adjustment strategy is used to adjust the exposure time, expressed as:

[0084]

[0085] in, Indicates the exposure adjustment step size. Represents the average gray value Compared with reference value The difference between them, the reference value of the average pixel grayscale. Within the 30-40 range.

[0086] like Figure 5 As shown, the position of the light spot is extracted using the centroid tracking method, and the centroid of the light spot ( , )for:

[0087]

[0088] Where 1≤k≤K, 1≤l≤L, and K and L are the width and height of the window in pixels, respectively; Representing pixels The x and y coordinates, Represents pixels The grayscale value; if , then let , The set grayscale threshold; if , then let .

[0089] If the light spot loses lock due to interference, the coarse tracking control unit controls the electrically switched filter to a visible / near-infrared broadband filter. The coarse tracking control unit controls the visible / near-infrared camera to adopt an image adaptive high exposure mode and increases the exposure time adjustment to quickly recover image tracking using the acquired target image.

[0090] Increase the adjusted exposure time :

[0091]

[0092] Furthermore, to eliminate the additional latency caused by image modality processing, this embodiment employs an inverse model feedforward PID control strategy, connecting an inverse model feedforward PID control module before the servo turntable, such as... Figure 6 As shown in Table 1, the controlled object is a servo turntable, and its dynamic characteristics are based on the LuGre friction model. The PID parameters for the image tracking stage and the spot tracking stage are shown in Table 1.

[0093] The transfer function of the PID controller is:

[0094]

[0095] in, This is the proportionality coefficient. The integral coefficient is... is the differential coefficient.

[0096] Table 1 PID parameters of the dual-mode tracking system

[0097] Image tracking stage 8 2 0.1 Spot tracking stage 8.2 3.5 0.1

[0098] Feedforward controller The regularized inverse model of the second-order low-pass filter is represented as follows:

[0099]

[0100] in, For the transfer function of the servo turntable, , The filter time constant is Let be the damping ratio, and s be the Laplace domain operator.

[0101] Using the method described in this embodiment, the link establishment time is significantly shortened and the capture performance is significantly improved: Experimental data shows that the dual-mode scanning method significantly reduces the average link establishment time from 2.8 seconds of the traditional method to 0.7 seconds, and the average image capture success rate is as high as 96.67%. The combined dual-mode capture probability is about 3.97% higher than that of the traditional single spot mode.

[0102] The method described in this embodiment significantly accelerates the recovery capability after a lock loss: the key innovation of this technology lies in the substantial optimization of the recovery process after a lock loss. Experiments show that after a lock loss caused by active interference, the re-establishment speed is significantly reduced from 2.8 seconds in the traditional method to 0.5 seconds, and the success rate is increased from 82% to 98%. This is mainly due to the image modality recapture mechanism, which avoids the drawback of traditional solutions requiring a time-consuming global scan, greatly improving the robustness and availability of the system.

[0103] In summary, the method described in this invention systematically solves the bottlenecks in speed, reliability, and continuity of traditional APT systems through dual-mode deep fusion and intelligent control, providing practical and feasible key technical support for building a next-generation highly reliable, highly available, and fast-response space laser communication system.

Claims

1. A dual-mode acquisition and tracking method, characterized in that, Based on a visible light / near-infrared camera and an electrically switched filter, it includes an image scanning stage, an image capture stage, an image tracking stage, and a spot tracking stage; (a) Image scanning stage The electrical switching filter is switched to a visible / near-infrared broadband filter, and the visible / near-infrared camera adopts an image adaptive high exposure mode; a rectangular spiral scan is performed in a pre-set area of ​​uncertain angle; Real-time detection of the weighted average gray value of the image within the field of view The system determines whether a beacon spot has appeared. If the weighted average gray value exceeds 210, it is determined that a beacon spot has been detected, and the process proceeds to the spot tracking stage. (ii) Image capture stage During image scanning, a preset target grayscale template is simultaneously identified and matched, and the correlation coefficient between the currently scanned image to be matched and the target grayscale template is calculated. When the correlation coefficient is ≤0.8, continue image scanning; When the correlation coefficient exceeds 0.8, the image is considered successfully captured; if the weighted average gray value of the currently scanned image exceeds 210, the beacon spot is detected and the process proceeds to the spot tracking stage; if the beacon spot is not detected, the process proceeds to the image tracking stage. (III) Image Tracking Stage The KCF tracking method is used to continuously track the captured image target; a window is opened at the target location in the image to detect light spots, and the average gray value within the window is calculated. ;like Once the beacon spot is detected, the process transitions to the spot tracking phase. (iv) Spot tracking stage The electronically switched filter is switched to a near-infrared narrowband filter. The visible light / near-infrared camera adopts a spot-adaptive low-exposure mode, activates the spot mode adaptive exposure adjustment mechanism, and uses the centroid tracking method to extract the spot position.

2. The dual-mode acquisition and tracking method according to claim 1, characterized in that, During the image scanning phase, the weighted average gray value of the image within the field of view. for: ； Where 1≤n≤N, and N is the number of image pixels in the field of view; , Let n be the weight of the nth pixel. Let be the grayscale value of the nth pixel in the image.

3. The dual-mode acquisition and tracking method according to claim 2, characterized in that, During the image scanning stage, the exposure time adjustment process in the adaptive exposure adjustment mechanism for visible / near-infrared camera image modes is represented as follows: ； in, To adjust the exposure time, This is the current exposure time. As an exposure factor, The weighted average gray value of the image within the current field of view. With setting reference grayscale value The difference between them.

4. The dual-mode acquisition and tracking method according to claim 1, characterized in that, During the image capture stage, a normalized cross-correlation method is used for target grayscale template identification and matching, and the correlation coefficient is... for: ； Where (x, y) is the position of the target grayscale template; T represents the target grayscale template, and I represents the current scanned image to be matched. Both have a width of w and a height of h, 1≤i≤w, 1≤j≤h; The average pixel grayscale value of the scanned image to be matched. The average pixel grayscale value of the target grayscale template.

5. The dual-mode acquisition and tracking method according to claim 1, characterized in that, During the spot tracking stage, the adaptive exposure adjustment mechanism for the spot mode is as follows: Calculate the average gray value of pixels within the window. ; An error-based nonlinear step size adjustment strategy is used to adjust the exposure time, expressed as: ； in, Indicates the exposure adjustment step size. Represents the average gray value Compared with reference value The difference between them.

6. The dual-mode acquisition and tracking method according to claim 1, characterized in that, During the spot tracking stage, the spot position is extracted using the centroid tracking method, and the spot centroid ( , )for: ； Where 1≤k≤K, 1≤l≤L, and K and L are the width and height of the window in pixels, respectively; Representing pixels The x and y coordinates, Represents pixels The grayscale value; if , then let , The set grayscale threshold; if , then let .

7. The dual-mode acquisition and tracking method according to claim 1, characterized in that, During the spot tracking stage, if the spot is lost due to interference, the electronically switched filter is switched to a visible / near-infrared broadband filter. The visible / near-infrared camera adopts an image adaptive high exposure mode and increases the exposure time, returning to the image tracking stage. Increase the adjusted exposure time : 。 8. The dual-mode acquisition and tracking method according to claim 1, characterized in that, An inverse model feedforward PID control strategy is introduced, and an inverse model feedforward PID control module, including a PID controller and a feedforward controller, is connected in front of the servo turntable. The transfer function of the PID controller is: ； in, This is the proportionality coefficient. The integral coefficient is... Here are the differential coefficients, and s is the Laplace field operator; Feedforward controller The regularized inverse model of the second-order low-pass filter is represented as follows: ； in, For the transfer function of the servo turntable, , The filter time constant is The damping ratio is denoted as .

Images