Photoelectric pod dynamic target recapturing method based on mean variance

Abstract

The invention discloses a photoelectric pod dynamic target recapturing method based on a mean variance, and relates to the technical field of photoelectric pods, and the method comprises the steps: carrying out the continuous monitoring and analysis of a target in a field of view, and immediately triggering a target recapturing program when the target is detected to be gradually amplified in a picture until the picture is completely black; the method comprises the following steps of: continuously monitoring a picture, dividing the picture into a plurality of sub-regions by using an image processing technology, and analyzing a pixel value of each region; according to the photoelectric pod dynamic target re-capturing method based on the mean variance, a re-capturing program can be rapidly triggered under the condition that the target is temporarily lost due to the fact that the target is too close and the picture is all black, the region where the target may move can be rapidly positioned by monitoring and analyzing the picture in real time and combining the mean and variance calculation, and the target dynamic target re-capturing efficiency is improved. Therefore, efficient recapture of the target is realized, the overall efficiency of the system is improved, and the cost and the maintenance cost are reduced.

Description

Technical Field

[0001] The invention relates to the technical field of optoelectronic pods, and in particular to a mean-variance-based optoelectronic pod dynamic target recapture method. Background Art

[0002] In a complex and ever-changing battlefield environment, the high mobility of the target poses a great challenge to the tracking of the optoelectronic pod. Especially when the distance between the target and the optoelectronic pod is too close, the imaging system may temporarily fail due to exceeding its designed working distance, resulting in a completely black screen and the loss of the target. Once this happens, quickly determining the target's moving direction and recapture the target becomes the key to improving the guidance effect. However, the existing target recapture technology usually has the problem of slow response speed, and it takes a long time to relock the target. This not only reduces the accuracy and effect of the missile's strike, but also affects the missile's survivability. Therefore, in order to address this problem, it is necessary to further study and optimize the target recapture technology to improve its response speed and locking accuracy. To this end, we propose a dynamic target recapture method for optoelectronic pods based on mean variance. Summary of the invention

[0003] In order to solve the above technical problems, a method for dynamic target recapture of an optoelectronic pod based on mean variance is provided. This technical solution solves the above problems.

[0004] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0005] A method for recapture of dynamic targets of an optoelectronic pod based on mean variance, comprising:

[0006] Continuously monitor and analyze the targets in the field of view. When the target is detected to be gradually enlarged in the picture until the picture is completely black, the target re-capture program is immediately triggered;

[0007] Monitor the screen continuously and use image processing technology to divide the screen into multiple sub-areas. Analyze the pixel values ​​of each area, calculate the mean and variance, and compare them with the mean and variance of the entire screen to identify specific areas that are significantly different from the overall screen.

[0008] Based on the obtained mean and variance data, the current position of the target and its future movement direction are predicted by comprehensive analysis based on statistical principles and the target's motion trajectory in previous frames;

[0009] Automatically adjust the search direction of the optoelectronic pod based on the inferred target movement direction;

[0010] When the target is successfully identified and located, the tracking procedure is immediately initiated to ensure that the target re-enters the monitoring range of the guidance system and the target tracking parameters are reset to adapt to the latest dynamics of the target.

[0011] Preferably, the target in the field of view is continuously monitored and analyzed, and when it is detected that the target is gradually enlarged in the picture until the picture is completely black, the target re-capture procedure is immediately triggered, which specifically includes:

[0012] In the initial stage of the guidance system, the target in the field of view is continuously monitored and analyzed by the image acquisition equipment;

[0013] If the target continues to grow and blocks the image acquisition device, causing the light to be blocked, making the entire screen black or the brightness drop sharply, the system determines whether the screen is close to being completely black;

[0014] Among them, the method for judging whether the screen is completely black is:

[0015] Assume there are l pixels in the picture, and the gray value of each pixel is G e , the screen mean M is calculated as:

[0016]

[0017] M is compared with a preset threshold value T. If M<T, the image is determined to be completely black.

[0018] Preferably, using image processing technology to divide the picture into multiple sub-areas specifically includes:

[0019] According to the size of the picture and specific needs, the picture is evenly divided into rectangular sub-areas. For a picture with a resolution of C×K, the picture is divided into a×b rectangular sub-areas of equal size, and the size of each sub-area is C / a×K / b;

[0020] Starting from the upper left corner of the picture, the entire picture is gradually traversed according to the determined sub-region size, the picture is divided into multiple sub-regions, and a unique identifier is assigned to each sub-region;

[0021] For each sub-region, the RGB color model is used to extract the values ​​of the red, green, and blue channels of all pixels in it;

[0022] Convert each pixel's color value to a grayscale value.

[0023] Preferably, the gray value conversion formula is:

[0024] Gray=0.299R+0.587G+0.114B

[0025] Where R, G, B represent the values ​​of the red, green, and blue channels respectively, and Gtay represents the grayscale value.

[0026] Preferably, the pixel values ​​of each region are analyzed to calculate the mean and variance by:

[0027] Based on extracting color values ​​and converting them into grayscale values;

[0028] Calculate the average grayscale value of all pixels in each sub-area to obtain the mean brightness;

[0029] The calculation formula for the brightness mean is:

[0030]

[0031] In the formula, μ represents the mean brightness, N represents the total number of sub-regions, i represents the i-th sub-region, x represents the pixel grayscale value, and x i Represents the pixel gray value x in the i-th sub-region;

[0032] For each sub-region, calculate the sum of the squares of the differences between the grayscale value of each pixel and the mean, and then divide it by the number of pixels to get the variance;

[0033] The calculation formula of brightness variance is:

[0034]

[0035] In the formula, σ 2 Represents variance.

[0036] Preferably, comparing with the mean and variance of the entire picture to identify a specific area having significant differences from the entire picture specifically includes:

[0037] Based on the calculated mean and variance, for each sub-region, its mean is compared with the mean of the entire picture, and the absolute value of the difference between the sub-region mean and the overall picture mean is calculated:

[0038] Diff mean =|μ sub -μ total |

[0039] In the formula, Diff mean represents the mean difference, μ sub represents the mean of the sub-region, μ total It represents the mean of the whole picture;

[0040] For each sub-region, its variance is compared with the variance of the entire picture, and the absolute value of the difference between the variance of the sub-region and the variance of the entire picture is calculated:

[0041]

[0042] In the formula, Diff varrepresents the variance difference, represents the variance of the sub-region, represents the variance of the overall picture;

[0043] Through multiple experiments and adjustments, a threshold is set to determine whether the difference between a sub-region and the overall picture is significant. If the mean difference and variance difference of the sub-region exceed the set threshold, then there is a significant difference between the sub-region and the overall picture, and it is marked as a specific region.

[0044] Preferably, predicting the current position of the target and its future moving direction specifically includes:

[0045] Collect the calculated mean and variance data of each sub-region, as well as the location information of the target in several previous frames, to form a historical data set;

[0046] Through statistical analysis methods, the correlation between the change of mean and variance and the target position is determined, and the change law of mean and variance of sub-areas under different target positions is observed;

[0047] For the position information of the target in the previous frame, the motion characteristics of the target are calculated, and the motion characteristics include: displacement vector, velocity and acceleration, and the motion trajectory of the target is fitted by linear regression;

[0048] A comprehensive analysis is performed based on the target position information reflected by the mean and variance data, as well as the target's motion trajectory characteristics. Based on the results of the comprehensive analysis, the current position of the target is predicted.

[0049] Preferably, the target's motion characteristic displacement vector, velocity and acceleration are calculated as follows:

[0050] Among them, the target displacement vector calculation formula is:

[0051]

[0052] In the formula, represents the displacement vector of the target, (x t ,y t ) represents the position coordinates of the target in the tth frame, (x t-1 ,y t-1 ) represents the position coordinates of the target in the t-1 frame;

[0053] The target speed calculation formula is:

[0054]

[0055] In the formula, represents the velocity vector of the target, Δt represents the time interval of target displacement;

[0056] The target acceleration calculation formula is:

[0057]

[0058] In the formula, represents the acceleration vector of the target, represents the speed of the target in frame t, Indicates the speed of the target at frame t-1.

[0059] Preferably, the calculation formula for predicting the current position of the target is:

[0060]

[0061] In the formula, (x pred ,y pred ) represents the current position prediction of the target, F represents the total number of different positions of the target, and p j represents the predicted probability of the jth position, (x j ,y j ) represents the coordinates of the jth position.

[0062] Preferably, the method steps for automatically adjusting the search direction of the optoelectronic pod according to the inferred target moving direction are:

[0063] The direction of the target's movement is inferred based on the mean and variance data and the target's previous movement trajectory;

[0064] According to the target moving direction vector, the angle that the optoelectronic pod needs to be adjusted is calculated using trigonometric functions. The formula for calculating the angle that the optoelectronic pod needs to be adjusted is:

[0065]

[0066] In the formula, θ represents the angle to be adjusted, arctan represents the inverse tangent function, dx represents the horizontal movement component, and dy represents the vertical movement component;

[0067] Based on the calculated angle, a command is sent to the optoelectronic pod motor control system to adjust the search direction of the optoelectronic pod toward the target movement direction.

[0068] Compared with the prior art, the present invention has the following beneficial effects:

[0069] The method for recapture of dynamic targets by an optoelectronic pod based on mean variance proposed by the present invention can quickly trigger a recapture program when the target is temporarily lost due to being too close and causing a completely black screen. By real-time monitoring and analyzing the screen and combining mean and variance calculations, the method can quickly locate the area where the target may move, thereby achieving efficient recapture of the target, greatly improving the response speed and strike accuracy of the guidance system. In a complex and changeable battlefield environment, the method can automatically adjust the search direction of the optoelectronic pod by accurately predicting the current position and future movement direction of the target, ensuring that the target re-enters the monitoring range of the guidance system, which not only improves the strike effect and accuracy of the missile and enhances the survivability of the missile, but also reduces unnecessary search and tracking processes through intelligent prediction and adjustment strategies, thereby optimizing resource utilization, not only improving the overall efficiency of the system, but also reducing costs and maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 is a flow chart of the method of the present invention;

[0071] Figure 2 Flowchart of a method for dividing a screen into multiple sub-areas;

[0072] Figure 3 A flow chart of a method for determining a target position prediction. DETAILED DESCRIPTION

[0073] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0074] Reference Figure 1 As shown, a method for recapture of dynamic targets of an optoelectronic pod based on mean variance includes:

[0075] Step 1: Analyze and detect targets

[0076] In the initial stage of the guidance system, the system analyzes and detects the target in the picture. When the target gradually grows in the field of view to completely cover the lens, it may cause the light to be blocked, making the entire screen black or the brightness drop sharply. If the picture is close to black, it usually means that the target is very close to the camera or beyond the range of the image acquisition system. At this time, the system will compare the mean of each sub-area with the preset threshold to determine whether the target is lost. If the mean of all or most of the sub-areas is lower than the threshold, the system determines that the target is lost.

[0077] The program will then analyze the image sequence in the video stream in real time, use advanced computer vision algorithms to identify abnormal situations where the screen turns black, and automatically issue a re-capture command. This process ensures that even if the target is temporarily lost due to the screen being completely black, the system can quickly re-capture it and achieve continuous and effective tracking of the target.

[0078] Step 2: Real-time monitoring and mean-variance calculation

[0079] In order to accurately monitor the target, the screen is divided into several sub-areas, which exist in the form of rectangles or squares and overlap with each other to cover the entire screen, ensuring that the target can be detected at any position. For each sub-area, the system calculates the mean and variance of the pixel values ​​in detail. The mean reflects the average pixel intensity in the area, while the variance measures the discrete degree of pixel value changes, that is, consistency or uniformity. High variance usually indicates that the pixel value changes are large, which may be caused by factors such as target movement, complex background texture or noise.

[0080] The system will traverse the pixels in each sub-region, extract the color value and convert it into grayscale value to simplify the calculation process. Then calculate the average of all the grayscale values ​​of the pixels in each sub-region to get the mean brightness. For each sub-region, calculate the sum of the squares of the difference between the grayscale value of each pixel and the mean, and then divide it by the number of pixels to get the variance.

[0081] Step 3: Determine the target's moving direction

[0082] As the video sequence is analyzed frame by frame, the system will continue to monitor the changes in the mean and variance of each sub-region. The movement of the target will affect the distribution of pixel values ​​in the sub-region where it is located, and thus affect the mean and variance. The movement of the target usually causes the mean of the sub-region in the new position to change, while in the area where the target moves out, the mean may decrease and the variance may decrease, because the movement of the target reduces the change in pixel values ​​in the area.

[0083] By comparing the mean and variance of each sub-region, the system can identify regions with feature changes that match the target's movement pattern. Regions where the variance is significantly reduced and the mean is changed are likely to be blanks or parts left behind by the target's movement.

[0084] Step 4: Adjust the search direction of the optoelectronic pod

[0085] Based on the target movement direction determined by the variance calculation results, the system will adjust the search direction of the optoelectronic pod accordingly, optimizing the search strategy based on the inferred target movement direction. By analyzing the movement direction and speed of the dark areas between adjacent frames, the system can predict the target's movement trajectory and adjust the pointing of the optoelectronic pod to align it with the target. Once the variance analysis determines the target's movement direction, the system then adjusts the optoelectronic pod's search direction to ensure that the target remains within the monitoring range.

[0086] Step 5: Determine your goals and track them again

[0087] Once the target location is determined by the system, it will immediately start the tracking procedure and reset the target tracking system. The system must respond quickly, adjust the electro-optical pod or sensor toward the target location, and activate the tracking mechanism to ensure effective tracking and monitoring of the target.

[0088] Continuously monitor and analyze the target in the field of view. When the target is detected to be gradually enlarged in the picture until the picture is completely black, the target re-capture procedure is immediately triggered. Specifically, it includes:

[0089] In the initial stage of the guidance system, the target in the field of view is continuously monitored and analyzed in depth through the image acquisition device. As time goes by, if the target continues to approach and grow in size within the field of view, its range gradually expands until it begins to block the field of view of the image acquisition device. When the target gets closer and causes serious obstruction to the image acquisition device, the light is greatly blocked, causing the brightness of the entire screen to change dramatically. At first, the screen may show an obvious darkening trend. As the target continues to approach and the degree of obstruction deepens, the brightness of the screen continues to decrease. During this process, the system will determine whether the picture is close to a completely black state through real-time monitoring and analysis of the screen brightness. Once the system determines that the picture is close to completely black, it will respond quickly and immediately trigger the target re-capture program to ensure that the target can be re-locked as soon as possible in the event of temporary loss, thereby ensuring the normal operation of the guidance system.

[0090] The use process of the present invention is: analyzing and detecting the target, real-time monitoring and mean variance calculation, judging the moving direction of the target, adjusting the search direction of the optoelectronic pod, and determining the target to track again.

[0091] In summary, the advantages of the present invention are: it can quickly trigger the recapture program when the target screen is completely black and temporarily lost, and combine the mean variance calculation to locate the possible target area, so as to achieve efficient recapture and improve the response speed and strike accuracy of the guidance system. In a complex battlefield environment, it can accurately predict the target position and movement direction, automatically adjust the search direction of the optoelectronic pod, ensure that the target returns to the monitoring range, improve the missile strike effect and accuracy, and enhance survivability. Through intelligent strategies, it can reduce unnecessary search and tracking, optimize resource utilization, improve the overall efficiency of the system, and reduce costs and maintenance costs.

[0092] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for recapture of dynamic targets of an optoelectronic pod based on mean variance, characterized in that: include: Continuously monitor and analyze the targets in the field of view. When the target is detected to be gradually enlarged in the picture until the picture is completely black, the target re-capture program is immediately triggered; Monitor the screen continuously and use image processing technology to divide the screen into multiple sub-areas. Analyze the pixel values ​​of each area, calculate the mean and variance, and compare them with the mean and variance of the entire screen to identify specific areas that are significantly different from the overall screen. Based on the obtained mean and variance data, the current position of the target and its future movement direction are predicted by comprehensive analysis based on statistical principles and the target's motion trajectory in previous frames; Automatically adjust the search direction of the optoelectronic pod based on the inferred target movement direction; When the target is successfully identified and located, the tracking procedure is immediately initiated to ensure that the target re-enters the monitoring range of the guidance system and the target tracking parameters are reset to adapt to the latest dynamics of the target.

2. The method for recapture of dynamic targets by an optoelectronic pod based on mean variance according to claim 1, characterized in that: Continuously monitor and analyze the target in the field of view. When the target is detected to be gradually enlarged in the picture until the picture is completely black, the target re-capture procedure is immediately triggered. Specifically, it includes: In the initial stage of the guidance system, the target in the field of view is continuously monitored and analyzed by the image acquisition equipment; If the target continues to grow and blocks the image acquisition device, causing the light to be blocked, making the entire screen black or the brightness drop sharply, the system determines whether the screen is close to being completely black; Among them, the method for judging whether the screen is completely black is: Assume there are l pixels in the picture, and the gray value of each pixel is G e , the screen mean M is calculated as: M is compared with a preset threshold value T. If M<T, the image is determined to be completely black.

3. The method for recapture of dynamic targets by an optoelectronic pod based on mean variance according to claim 1, characterized in that: Using image processing technology to divide the picture into multiple sub-areas specifically includes: According to the size of the picture and specific needs, the picture is evenly divided into rectangular sub-areas. For a picture with a resolution of C×K, the picture is divided into a×b rectangular sub-areas of equal size, and the size of each sub-area is C / a×K / b; Starting from the upper left corner of the picture, the entire picture is gradually traversed according to the determined sub-region size, the picture is divided into multiple sub-regions, and a unique identifier is assigned to each sub-region; For each sub-region, the RGB color model is used to extract the values ​​of the red, green, and blue channels of all pixels in it; Convert each pixel's color value to a grayscale value.

4. The method for recapture of dynamic targets of an optoelectronic pod based on mean variance according to claim 1, characterized in that: The gray value conversion formula is: Gray=0.299R+0.587G+0.114B Where R, G, B represent the values ​​of the red, green, and blue channels respectively, and Gtay represents the grayscale value.

5. The method for recapture of dynamic targets by an optoelectronic pod based on mean variance according to claim 1, characterized in that: The pixel values ​​of each area are analyzed and the method to calculate the mean and variance is as follows: Based on extracting color values ​​and converting them into grayscale values; Calculate the average grayscale value of all pixels in each sub-area to obtain the mean brightness; The calculation formula for the brightness mean is: In the formula, μ represents the mean brightness, N represents the total number of sub-regions, i represents the i-th sub-region, x represents the pixel grayscale value, and x i Represents the pixel gray value x in the i-th sub-region; For each sub-region, calculate the sum of the squares of the differences between the grayscale value of each pixel and the mean, and then divide it by the number of pixels to get the variance; The calculation formula of brightness variance is: In the formula, σ 2 Represents variance.

6. The method for recapture of dynamic targets by an optoelectronic pod based on mean variance according to claim 1, characterized in that: By comparing the mean and variance of the entire picture, specific areas that are significantly different from the overall picture are identified, including: Based on the calculated mean and variance, for each sub-region, its mean is compared with the mean of the entire picture, and the absolute value of the difference between the sub-region mean and the overall picture mean is calculated: Diff mean =|μ sub -m total | In the formula, Diff mean represents the mean difference, μ sub represents the mean of the sub-region, μ total It represents the mean of the whole picture; For each sub-region, its variance is compared with the variance of the entire picture, and the absolute value of the difference between the variance of the sub-region and the variance of the entire picture is calculated: In the formula, Diff var represents the variance difference, represents the variance of the sub-region, represents the variance of the overall picture; Through multiple experiments and adjustments, a threshold is set to determine whether the difference between a sub-region and the overall picture is significant. If the mean difference and variance difference of the sub-region exceed the set threshold, then there is a significant difference between the sub-region and the overall picture, and it is marked as a specific region.

7. The method for recapture of dynamic targets by an optoelectronic pod based on mean variance according to claim 1, characterized in that: Predicting the current position of the target and its future movement direction specifically includes: Collect the calculated mean and variance data of each sub-region, as well as the location information of the target in several previous frames, to form a historical data set; Through statistical analysis methods, the correlation between the change of mean and variance and the target position is determined, and the change law of mean and variance of sub-areas under different target positions is observed; For the position information of the target in the previous frame, the motion characteristics of the target are calculated, and the motion characteristics include: displacement vector, velocity and acceleration, and the motion trajectory of the target is fitted by linear regression; A comprehensive analysis is performed based on the target position information reflected by the mean and variance data, as well as the target's motion trajectory characteristics. Based on the results of the comprehensive analysis, the current position of the target is predicted.

8. The method for recapture of dynamic targets by an optoelectronic pod based on mean variance according to claim 6 is characterized in that: The target's motion characteristic displacement vector, velocity and acceleration are calculated as follows: Among them, the target displacement vector calculation formula is: In the formula, represents the displacement vector of the target, (x t ,y t ) represents the position coordinates of the target in the tth frame, (x t-1 ,y t-1 ) represents the position coordinates of the target in the t-1 frame; The target speed calculation formula is: In the formula, represents the velocity vector of the target, Δt represents the time interval of target displacement; The target acceleration calculation formula is: In the formula, represents the acceleration vector of the target, represents the speed of the target in frame t, Indicates the speed of the target at frame t-1.

9. The method for recapture of dynamic targets by an optoelectronic pod based on mean variance according to claim 6, characterized in that: The calculation formula for predicting the current position of the target is: In the formula, (x pred ,y pred ) represents the current position prediction of the target, F represents the total number of different positions of the target, and p j represents the predicted probability of the jth position, (x j ,y j ) represents the coordinates of the jth position.

10. The method for recapture of dynamic targets of an optoelectronic pod based on mean variance according to claim 1, characterized in that: The method steps for automatically adjusting the search direction of the optoelectronic pod according to the inferred target movement direction are as follows: The direction of the target's movement is inferred based on the mean and variance data and the target's previous movement trajectory; According to the target moving direction vector, the angle that the optoelectronic pod needs to be adjusted is calculated using trigonometric functions. The formula for calculating the angle that the optoelectronic pod needs to be adjusted is: In the formula, θ represents the angle to be adjusted, arctan represents the inverse tangent function, dx represents the horizontal movement component, and dy represents the vertical movement component; Based on the calculated angle, a command is sent to the optoelectronic pod motor control system to adjust the search direction of the optoelectronic pod toward the target movement direction.

Images