Theodolite stable tracking method based on multi-frame information fusion

Abstract

The invention relates to a theodolite stable tracking method based on multi-frame information fusion, belongs to the technical field of photoelectric theodolites, and solves the problems of insufficient stability and accuracy of the theodolites in a target tracking process. The method comprises the following steps: extracting a gate image by using a centroid tracking method; carrying out stable tracking state judgment based on multi-frame information fusion; in a tracking to-be-evaluated state in a stable tracking state, performing target trajectory prediction by using a Kalman filter based on a comprehensive angle value; and converting the predicted comprehensive angle value into a predicted target miss distance, and comparing a deviation value between the predicted target miss distance and an actual miss distance with a threshold value so as to perform adaptive correction on the miss distance. According to the method, the stable tracking state of the theodolite is judged through multi-frame information fusion, trajectory prediction is performed in combination with the Kalman filter, and the method has the advantages that the tracking robustness and accuracy are enhanced, the miss distance is adaptively corrected, and the tracking performance of the theodolite in a complex environment is improved.

Description

Technical Field

[0001] The invention relates to the technical field of photoelectric theodolites, and in particular to a theodolite stable tracking method based on multi-frame information fusion. Background Art

[0002] Photoelectric theodolites are mainly used for trajectory measurement and live recording of ballistic targets such as missiles and launch vehicles. The observed target is kept in the center of the field of view through the coordination between the target miss distance extracted in real time by the image processing subsystem and the drive control of the servo subsystem. However, when tracking targets under a complex sky background, it is easily disturbed by dynamic processes such as cloud movement, bird flight, and target separation, resulting in abnormal miss distance extraction and gate mutation, which affects the automatic tracking and measurement performance of the system and causes the target to leave the observation field of view.

[0003] The current photoelectric theodolite is mainly composed of an optical subsystem (visible, infrared, etc.), a tracking frame, a servo subsystem, an operation control subsystem, an image tracking subsystem, an image storage subsystem, a video interpretation subsystem, a time system terminal, etc., which are used to complete the tracking, measurement and live recording of flight targets such as missiles and launch vehicles with predetermined trajectories. Among them, the servo subsystem and the image tracking subsystem are key components for completing the target tracking process. In the process of tracking high-speed moving targets, the area pointed to by the photoelectric theodolite will form a projection image on the target surface of the detection device. The image tracking subsystem extracts the target miss amount in real time and sends it to the servo control system as the basis for controlling the movement of the theodolite. Then the servo subsystem controls the turntable to track the target and obtains the next frame of image at the same time, so as to form a closed-loop automatic tracking control. However, the image-based automatic tracking strategy is inevitably affected by the accuracy of the tracking algorithm. When the target is located in a complex sky background, the accuracy of the tracking algorithm is easily disturbed by dynamic processes such as cloud movement, bird flight and target separation, resulting in abnormal miss amount extraction, causing the target to leave the field of view and tracking measurement failure. Therefore, how to improve the stability and accuracy of target miss distance extraction has always been a key issue that cannot be ignored in the tracking process of photoelectric theodolite.

[0004] There are currently three mainstream solutions to the problem of insufficient stability and accuracy of photoelectric theodolites during target tracking, namely template matching, trajectory prediction, and deep learning. However, each of these methods has certain limitations:

[0005] Template matching method: When tracking a target, the performance of the template matching method is significantly dependent on the selection of the initial template, and is highly sensitive to the variation of the target's appearance and the dynamics of the background. When the target undergoes transformations such as scaling, rotation, and occlusion, the accuracy of template matching is often negatively affected.

[0006] Trajectory prediction method: This method analyzes the miss distance data of the target in the previous few frames to construct a prediction model for calculating the position of the target in the next frame. However, this strategy does not fully consider the accuracy of the miss distance data in the previous few frames. If there are errors in the data of the previous few frames, this incorrect information will be directly transmitted to the prediction model, thereby affecting the accuracy of trajectory prediction.

[0007] Deep learning method: This technical route relies on the construction of a large-scale dataset for network training, and its performance is greatly affected by the texture and feature information of the target. Especially when dealing with small infrared targets, due to the lack of sufficient texture feature information, the detection accuracy often fails to meet the requirements. In addition, the computational complexity of deep learning algorithms is relatively high, resulting in challenges in achieving high-frame-rate real-time processing. Summary of the Invention

[0008] In view of the above problems, the present invention proposes a theodolite stable tracking method based on multi-frame information fusion. Based on the traditional theodolite centroid tracking algorithm, this method performs multi-modal information statistics with multi-frame association, and proposes a tracking state determination method based on the fusion of target gray level and position information to evaluate the credibility of the correct extraction of the target miss distance. At the same time, a Kalman filter target trajectory prediction method based on the comprehensive angle value is proposed, which converts the predicted comprehensive angle value into a predicted target miss distance, and uses the predicted miss distance to correct the true miss distance, finally achieving the stable and accurate output of the target miss distance.

[0009] The technical solution adopted by the present invention is as follows:

[0010] A theodolite stable tracking method based on multi-frame information fusion, the method comprising the following steps:

[0011] Step 1: Extract the gate image using the centroid tracking method;

[0012] Step 2: Determine the stable tracking state based on multi-frame information fusion, specifically including the steps:

[0013] Step 2.1: Calculate the target gray level evaluation value of each frame of image, and continuously count the average value of the target gray level evaluation values of the

[0014] frames of images, and calculate the corresponding comprehensive target gray level evaluation value according to this average value;

[0015] Step 2.2: Calculate the intersection over union of the target detection gate of the current frame image and the target detection gate of the previous frame image; Based on the intersection over union of frame images, the target gray evaluation value, and the comprehensive target gray evaluation value, the tracking state of the theodolite is classified into one of the stable tracking state, the tracking to-be-evaluated state, and the unstable tracking state;

[0016] Step 3: In the tracking to-be-evaluated state under the stable tracking state, perform Kalman filter target trajectory prediction based on the comprehensive angle value, specifically including the steps:

[0017] Step 3.1: Calculate the comprehensive angle value of the target according to the azimuth angle, elevation angle when the theodolite collects images, and the miss distance of the target obtained after target detection of the image;

[0018] Step 3.2: Construct a Kalman filter and use the Kalman filter to predict the trajectory state of the target in the next frame image, where the trajectory state includes the predicted comprehensive angle value of the target;

[0019] Step 4: Convert the predicted comprehensive angle value into a predicted target miss distance, and by comparing the deviation value between the predicted target miss distance and the actual miss distance detected in the current frame image with a threshold, output the adaptively corrected miss distance.

[0020] The theodolite stable tracking method based on multi-frame information fusion proposed by the present invention realizes the determination of the stable tracking state of the theodolite through multi-frame information fusion, and combines Kalman filtering for trajectory prediction, having the following advantages: enhancing the robustness and accuracy of the theodolite tracking, and optimizing the tracking strategy through state classification; Kalman filtering improves the prediction stability and accuracy, realizes the adaptive correction of the miss distance, ensures the real-time performance and anti-interference ability, and improves the tracking performance of the optoelectronic theodolite in a complex environment. Brief Description of the Drawings

[0021] Figure 1 is a flowchart of a theodolite stable tracking method based on multi-frame information fusion according to an embodiment of the present invention;

[0022] Figure 2 is a schematic diagram of the comprehensive angle value. Detailed Embodiment

[0023] Next, the technical solution of the present invention will be described in detail in conjunction with the accompanying drawings and preferred embodiments.

[0024] The automatic tracking process of the photoelectric theodolite mainly includes: the image processing system captures the full-field target based on the received current frame image. If the capture is successful, it draws a gate and sends the miss distance information of the current target to the servo control system. The servo control system drives the theodolite to rotate for real-time target tracking according to the received miss distance information and updates the next frame image. If the capture fails, the servo control system performs guided tracking according to the external guidance data. At this time, the image processing system enters the capture state and continues to capture the target of interest, and so on to form a closed-loop automatic tracking.

[0025] As Figure 1 shown, this embodiment provides a steady tracking method for the theodolite based on multi-frame information fusion, and this method includes the following steps 1 to 4.

[0026] Step 1: Use the classic centroid tracking method to achieve the extraction of the gate image.

[0027] The centroid tracking method is generally implemented based on the principle of threshold segmentation, so the selection of the threshold is crucial. This embodiment uses a threshold segmentation method based on gray-scale statistics to determine the threshold , and the main idea of the threshold segmentation method based on gray-scale statistics is expressed by the following formula:

[0028] ;

[0029] ;

[0030] Among them, , are the length and width of the processed image respectively; is the average gray value of the whole image; is the gray standard deviation of the whole image; is a constant. Generally, takes 3; is an adjustable threshold deviation value; is the gray value of the target point at the th row and the th column of the original image.

[0031] After the threshold is determined, the pixels with gray values greater than the threshold are divided into the target area, and the pixels with gray values less than the threshold are divided into the background area, which is expressed by the following formula:

[0032] ;

[0033] Among them, represents the th row and the The gray value of the column pixel, represents the gray value of the pixel at the th row and th column of the original image.

[0034] For the set of all target pixels obtained according to the above classification decision, calculate the centroid coordinates of the target according to the following formula:

[0035] ;

[0036] ;

[0037] where: is the gray value of the pixel at the th row and th column of the image after threshold segmentation, is the centroid row coordinate of the target, is the centroid column coordinate of the target.

[0038] Perform gated image extraction based on the segmented target region and the centroid coordinates of the target. The specific operation is as follows: Extract the minimum bounding rectangle of the target region, with the upper left corner coordinate being , and the lower right corner coordinate . Then the width and height of the gated image are respectively:

[0039] ;

[0040] ;

[0041] where, is generally a constant, and the coordinate of the upper left corner of the gated image is .

[0042] Step 2: Determine the stable tracking state based on multi-frame information fusion, specifically including the following steps 2.1 - step 2.3.

[0043] Step 2.1: Calculate the target gray evaluation value and the comprehensive target gray evaluation value.

[0044] Perform gray calculation on the gated image detected in each frame to obtain the target gray evaluation value. If the detection fails, calculate the target gray evaluation value for the center window of half of the image. The calculation formula for the target gray evaluation value is as follows:

[0045] ;

[0046] where, is the gray mean value of the current gated image, is the gray maximum value of the current gated image.

[0047] Continuous statistics The average value of the target gray-scale evaluation value of the frame gating image, which is a gray-scale information evaluation index for judging whether the target in the next frame is successfully detected. That is, according to The average value of the target gray-scale evaluation value of the frame gating image, the comprehensive target gray-scale evaluation value is calculated The calculation formula is as follows:

[0048] ;

[0049] Among them, , is an adjustable gray-scale threshold.

[0050] Step 2.2: Calculate the intersection over union.

[0051] Calculate the intersection over union value of the target detection gate of the current frame image and the target detection gate of the previous frame image, which is a position information evaluation index for judging whether the target tracking is stable. The calculation formula of the intersection over union is as follows:

[0052] ;

[0053] Among them, and respectively represent the target detection gate of the current frame and the target detection gate of the previous frame, represents the intersection area of the two rectangular gate frames, while represents the union area of the two rectangular gate frames; represents the area of the region.

[0054] Step 2.3: Determine the stable tracking state.

[0055] Judge whether the current theodolite is in a stable tracking state. If the current is not in a stable tracking state, then judge whether it meets the conditions of three consecutive frames of images and . If it meets the conditions, judge that the current theodolite is in a stable tracking state. If it does not meet the conditions, judge that the current theodolite is in an unstable tracking state and determine that the target is lost. If the current theodolite is in a stable tracking state, then execute the following state determination strategy:

[0056] Statistically, the intersection over union of the continuous frames of images before the current moment, and each frame meets and . When this is the case, it is considered that the target gate extraction is stable and the confidence level of correct target extraction is relatively high, and the miss distance of the target can be output. At this time, it can enter the stable tracking state. Generally, can take 3;

[0057] In the stable tracking state, when When it is in this situation, it is considered that the accuracy of the target gate extraction needs to be evaluated at this time, and there is a risk of interference from other targets. At this time, it enters the state of tracking to be evaluated;

[0058] In the stable tracking state, if it is considered that the target extraction gate jumps severely, and the confidence level of correct target extraction is low. It is determined that the target is lost, and at this time, it enters the unstable tracking state.

[0059] Step 3: In the state of tracking to be evaluated in the stable tracking state, based on the Kalman filter target trajectory prediction of the comprehensive angle value, specifically including the following steps 3.1 - step 3.2.

[0060] Step 3.1: Calculate the comprehensive angle value.

[0061] As Figure 2 shown, the azimuth angle and elevation angle when the theodolite collects images are and respectively. In the imaging coordinate system with the center of the field of view being P, the tracking algorithm performs target capture. At this time, the miss distance of target detection is and , then the azimuth angle offset of the target relative to the center of the field of view is , and the elevation angle offset is . At this time, the comprehensive angle value of the target is:

[0062] ;

[0063] ;

[0064] Among them, is the pixel size, is the system focal length, is the field of view angle occupied by one pixel.

[0065] Step 3.2: Construct a Kalman filter and use the Kalman filter to predict the trajectory state of the target in the next frame of image, and this trajectory state includes the predicted comprehensive angle value of the target.

[0066] Generally, the flight trajectory of the target in the task is smooth. Therefore, the state estimation and trajectory prediction of the comprehensive angle value of the target in the next frame of image are carried out to assist the target tracking strategy of the theodolite. The present invention uses a Kalman filter with a uniform acceleration and linear observation model to predict the trajectory state of the target in the next frame of image. Therefore, the state vector of the target at time can be modeled as:

[0067] ;

[0068] Among them, , respectively represent the comprehensive angle values of the target at moment, , respectively are the rate of change of the comprehensive angle value of the target at moment.

[0069] In step 3.2, when using the Kalman filter to predict the trajectory state of the target in the next frame image, the following steps are included:

[0070] Step 3.2.1: Determine whether the Kalman filter is in a stable prediction state. If so, execute step 3.2.2; otherwise, execute step 3.2.4.

[0071] In the stable tracking state, the comprehensive angle values predicted by the Kalman filter for the trajectory at the current moment are and . The true comprehensive angle values detected for the target in the current frame at the current moment are and , then define the comprehensive angle prediction deviation value as:

[0072] ;

[0073] Continuously count the comprehensive angle prediction deviation values for frames. If frames all satisfy , it is considered that the Kalman filter is in a stable prediction state. Generally, .

[0074] Step 3.2.2: Determine whether the theodolite is in a state to be evaluated for tracking. If so, execute step 3.2.3; otherwise, execute step 3.2.4.

[0075] Step 3.2.3: Predict the trajectory state of the target in the current frame image, including the predicted comprehensive angle value of the target.

[0076] In the prediction stage, use the time update equation to predict the target state in the current frame image by using the comprehensive angle value of the target in the previous frame image, where the time update equation is:

[0077] ;

[0078] ;

[0079] Among them, is the state transition matrix, is the current moment measurement error covariance matrix predicted based on the covariance of the previous moment, is the system noise, The measurement error covariance matrix of the optimal estimate for the previous frame.

[0080] Step 3.2.4: Update the Kalman filter.

[0081] In the update phase, the comprehensive angle value of the target in the current frame image is used to correct the predicted state of the Kalman filter system. After correction, the miss distance is output, and the next iteration is entered to predict the comprehensive angle value of the target in the next frame image. The state update equation is as follows:

[0082] ;

[0083] ;

[0084] ;

[0085] Where, is the Kalman gain, is the measurement error, is the observation matrix, is the optimal estimated value of the target trajectory state, is the predicted value of the target trajectory state, is the actual measured value of the target trajectory state, is the identity matrix, is the measurement error covariance matrix.

[0086] Step 4: Convert the predicted comprehensive angle value into the predicted target miss distance and output the miss distance after adaptive correction.

[0087] If it is currently in the state of tracking to be evaluated, the Kalman filter is used to predict the trajectory of the current frame. The predicted comprehensive angle value of the target is and , the azimuth angle and the elevation angle of the current optical measurement device. At this time, the predicted target miss distances and are respectively:

[0088] ;

[0089] ;

[0090] The actual target miss distances extracted at this time are and . If the deviation value between the predicted target miss distance and the actual miss distance detected in the current frame image is less than the threshold, that is, , it is considered that the target miss distance detected in the current frame is relatively consistent with the predicted miss distance, and the actual miss distance of the target is output and If the deviation value between the predicted target miss distance and the actual miss distance detected in the current frame image is greater than the threshold, that is , it is considered that the target miss distance detected in the current frame is quite different from the predicted target miss distance, and there is a high risk of detection error in the actual miss distance of the target detected in the current frame. Then the predicted target miss distance is output and to achieve adaptive correction of the output miss distance.

[0091] The present invention proposes a theodolite stable tracking method based on multi-frame information fusion. Through multi-frame correlation statistics, the stable tracking state determination of the theodolite is realized by using the comprehensive target gray evaluation value of multiple frames and the intersection over union of the target gate image. The tracking state of the theodolite is divided into a stable tracking state, a tracking to-be-evaluated state, and an unstable tracking state to evaluate the correctness and credibility of the extraction of the target miss distance.

[0092] In the stable tracking state, a Kalman filter target trajectory prediction method based on the comprehensive angle value is constructed. This method proposes a target trajectory stable prediction state determination strategy. In the stable prediction state, the miss distance extracted in the tracking to-be-evaluated state is analyzed and determined. The predicted comprehensive angle value is converted into a predicted miss distance, and through comparison with the threshold of the actual miss distance, the adaptive correction of the output miss distance is realized.

[0093] The present invention has the following advantages:

[0094] 1) Compared with the template matching method: The method of the present invention does not depend on the selection of the template and will not be affected when there are variations in the target appearance;

[0095] 2) Compared with the trajectory prediction method: The Kalman filter constructed by the method of the present invention is in the stable tracking state and predicts the target trajectory in the stable prediction state, with higher prediction accuracy and more guarantee;

[0096] 3) Compared with the deep learning method: The method of the present invention does not require building a large amount of data sets and does not depend on the texture feature information of the target, with lower computational complexity and strong applicability.

[0097] The theodolite stable tracking method based on multi-frame information fusion proposed by the present invention realizes the stable tracking state determination through multi-frame information fusion, and combines the Kalman filter for trajectory prediction, having the following advantages: enhancing the robustness and accuracy of the theodolite tracking, optimizing the tracking strategy through state classification; the Kalman filter improves the prediction stability and accuracy, realizes the adaptive correction of the miss distance, ensures the real-time performance and anti-interference ability, and improves the tracking performance of the optoelectronic theodolite in complex environments.

[0098] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0099] The above-described embodiments only express several implementation manners of the present invention, and the description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several deformations and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the invention patent should be subject to the appended claims.

Claims

1. A theodolite stable tracking method based on multi-frame information fusion, characterized in that Including the following steps: Step 1: Extract the gate image using the centroid tracking method; Step 2: Determine the stable tracking state based on multi-frame information fusion, specifically including the steps: Step 2.1: Calculate the target gray level evaluation value of each frame of image, and continuously count the average value of the target gray level evaluation values of the frames of images, and calculate the corresponding comprehensive target gray level evaluation value according to this average value; Step 2.2: Calculate the intersection over union of the target detection gate in the current frame image and the target detection gate in the previous frame image; Step 2.3: Determine whether the current theodolite is in a stable tracking state. If so, classify the tracking state of the theodolite into one of the stable tracking state, tracking to be evaluated state, and unstable tracking state according to the intersection over union and target gray evaluation value, and comprehensive target gray evaluation value of continuously counted frame images; Step 3: Track the state to be evaluated in the stable tracking state, and predict the target trajectory based on the Kalman filter of the comprehensive angle value, specifically including the steps: Step 3.1: Calculate the comprehensive angle value of the target according to the azimuth angle, elevation angle when the theodolite collects the image, and the miss distance of the target obtained after target detection of the image; Step 3.2: Construct a Kalman filter, and use the Kalman filter to predict the trajectory state of the target in the next frame image, where the trajectory state includes the predicted comprehensive angle value of the target; Step 4: Convert the predicted comprehensive angle value into a predicted target miss distance, compare the deviation value between the predicted target miss distance and the actual miss distance detected in the current frame image with a threshold, and output the adaptively corrected miss distance.

2. The theodolite stable tracking method based on multi-frame information fusion according to claim 1, wherein In step 1, the calculation formula of the threshold in the centroid tracking method is as follows: ； ； Among them, and are the length and width of the processed image respectively; is the average gray value of the whole image; is the standard deviation of gray value of the whole image; is a constant; is an adjustable threshold deviation value; is that the target point is located at the gray value of the pixel in the th row and the th column of the original image.

3. A method for stable tracking of a theodolite based on multi-frame information fusion according to claim 1 or 2, characterized in that, In Step 2.1, the calculation formula of the target gray scale evaluation value is as follows: ； Among them, is the gray-scale average value of the current gate image, is the maximum gray-scale value of the current gate image; The calculation formula of the comprehensive target gray scale evaluation value is as follows: ； Among them, , is an adjustable gray-scale threshold value.

4. A method for stable tracking of a theodolite based on multi-frame information fusion according to claim 1 or 2, characterized in that In Step 2.2, the calculation formula of the intersection over union is as follows; ； Among them, and respectively represent the target detection gate of the current frame image and the target detection gate of the previous frame image, represents the intersection area of the two rectangular wave door frames, represents the union area of the two rectangular wave door frames; represents the area of the region.

5. A method for stable tracking of a theodolite based on multi-frame information fusion according to claim 1 or 2, characterized in that, In Step 3.1, the calculation formula of the comprehensive angle value of the target is as follows: ； ； Among them, is the pixel size, is the system focal length, is the field of view angle occupied by one pixel; and are the azimuth angle and elevation angle of the theodolite respectively; and are the azimuth angle offset and elevation angle offset of the target relative to the center of the field of view respectively.

6. A method for stable tracking of a theodolite based on multi-frame information fusion according to claim 1 or 2, characterized in that, In Step 3.2, when using the Kalman filter to predict the trajectory state of the target in the next frame image, it includes the following steps: Step 3.2.1: Judge whether the Kalman filter is in a stable prediction state. If so, execute Step 3.2.2; otherwise, execute Step 3.2.4; Step 3.2.2: Judge whether the theodolite is in the state to be evaluated for tracking. If so, execute Step 3.2.3; otherwise, execute Step 3.2.4; Step 3.2.3: Prediction stage: Predict the trajectory state of the target in the current frame image by using the comprehensive angle value of the target in the previous frame image through the time update equation; Step 3.2.4: Update stage: Correct the prediction state of the system by using the comprehensive angle value of the target in the current frame image, and enter the next iteration to realize the prediction of the comprehensive angle value of the target in the next frame image.

7. A method for stable tracking of a theodolite based on multi-frame information fusion according to claim 6, characterized in that, In Step 3.2.1, the following method is used to judge whether the Kalman filter is in a stable prediction state: Define the comprehensive angle prediction deviation value as: ； Among them, and are respectively the true comprehensive angle values of the current frame object detection, and are respectively the comprehensive angle values obtained by trajectory prediction through the Kalman filter at the current moment; Continuous statistics The comprehensive angular prediction deviation value of the frame image. If it is continuous of the frame images all meet the conditions: , it is determined that the Kalman filter is in a stable prediction state.

8. A method for stable tracking of a theodolite based on multi-frame information fusion according to claim 6, characterized in that, The time update equation in Step 3.2.3 is: ； ； Among them, is the optimal estimated value of the target trajectory state, is the state transition matrix, is the measurement error covariance matrix at the current moment predicted based on the covariance at the previous moment, is the system noise, is the measurement error covariance matrix of the optimal estimation of the previous frame.

9. A method for stable tracking of a theodolite based on multi-frame information fusion according to claim 6, characterized in that, The state update equation in Step 3.2.4 is: ； ； ； wherein, is the Kalman gain, is the measurement error, is the observation matrix, is the optimal estimated value of the target's trajectory state, is the predicted value of the target trajectory state, is the actual measured value of the target trajectory state, is the identity matrix, is the measurement error covariance matrix.

10. A method for stable tracking of a theodolite based on multi-frame information fusion according to claim 1 or 2, characterized in that, In Step 4, the predicted comprehensive angle value is converted into a predicted target miss distance through the following formula: ； ； Among them, is the pixel size, is the system focal length, and is the predicted comprehensive angle value, and are the azimuth angle and elevation angle of the current theodolite; If the deviation value between the predicted target miss amount and the actual miss amount detected in the current frame image is less than the threshold , then output the actual target miss amount and ; If the deviation value between the predicted target miss amount and the actual miss amount detected in the current frame image is greater than the threshold , then output the predicted target miss amount and .

Images