Saring-mode photoelectric detection searching and tracking calculation method and system

Abstract

The invention relates to the technical field of photoelectric detection searching and tracking, and discloses a staring-mode photoelectric detection searching and tracking calculation method and a staring-mode photoelectric detection searching and tracking calculation system. The method comprises the following steps: acquiring a plurality of shot continuous images, and performing feature extraction on the images to obtain a feature value of an image gradient change rule; calculating a frame difference for the characteristic values of the adjacent images, constructing a normal distribution model, calculating a standard deviation, determining an image threshold value based on a linear function of the standard deviation, re-acquiring the images, combining the image threshold value to separate a moving point, and recording a deviation value; performing convolution calculation on the deviation values of all the moving points to obtain deviation degrees, and selecting the moving point corresponding to the maximum deviation degree as a tracking target; windowing is carried out based on the coordinate position of the tracking target in combination with the preset distance, then a new moving point is repeatedly obtained, whether the moving point is within the windowing range or not is judged, if not, the tracking target is updated, and continuous tracking of the tracking target is completed. The problem that imaging is fuzzy in the target searching and tracking process of existing photoelectric detection searching and tracking is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of optoelectronic detection, search and tracking, and particularly relates to a calculation method for optoelectronic detection, search and tracking in a staring mode. Background Art

[0002] Optoelectronic detection, search and tracking is a technology that uses optoelectronic detectors to search for and track targets, and is widely used in the fields of science and technology and security. Currently, optoelectronic detection, search and tracking mainly adopts a servo motion mode to scan the target area, usually including an optoelectronic detector, an image processing device and a servo system. By processing the image, a dynamic target is analyzed, and the servo system is controlled to drive the optoelectronic detector and the image processing device to follow the target in real time. To achieve the detection, search and tracking function, programmable array logic, digital signal processing and an image processing host are generally used to analyze the image frame by frame in real time and control the action of the servo system.

[0003] Since the detected target may move at high speed, optoelectronic detection, search and tracking often adopts a high-power servo system to ensure its tracking speed, so as to ensure that the load axis is always directed towards the target position, so as to ensure that subsequent possible disposal measures can be applied to the accurate target direction. Generally, optoelectronic detection, search and tracking does not need to apply disposal measures, and only needs to accurately record the behavior of the target for accurate recording, providing a factual basis for subsequent processing; in most cases where the target needs to be maneuvered and disposed at any time, the search and tracking technology has developed into "observation and aiming separation" or "observation and aiming follow-up". As a servo system with maneuverability, in the optoelectronic detection, search and tracking technology, the imaging of the detector is often blurred due to the rapid movement with excessive maneuverability, resulting in a loss of overall performance. Summary of the Invention

[0004] The present invention provides a calculation method and system for optoelectronic detection, search and tracking in a staring mode to solve the problem of blurred imaging in the process of searching and tracking a target by the existing optoelectronic detection, search and tracking.

[0005] To achieve the above object, the present invention is realized through the following technical solutions: In a first aspect, the present invention provides a calculation method for optoelectronic detection, search and tracking in a staring mode, which is applied to an optoelectronic detection device composed of a single camera and an image processing host, and includes the following steps: Step 1: Obtain a plurality of consecutive images captured by the optoelectronic detection device, extract features from all the images, and obtain the feature values of the gradient change law of all the images; Step 2: Calculate the frame difference of the feature values of all adjacent images and construct a normal distribution model, calculate the standard deviation of the variance value of the normal distribution model, determine an image threshold based on a linear function of the standard deviation, re-obtain the images captured by the optoelectronic detection device, and separate the moving points in combination with the image threshold and record the deviation value; Step 3: Perform convolution calculation on the deviation values of all moving points to obtain the deviation degree, and select the moving point corresponding to the maximum deviation degree as the tracking target; Step 4: Perform windowing based on the coordinate position of the tracking target in combination with a preset distance, and then repeat Steps 1 to 3 to obtain new moving points and determine whether they are within the window range. If they are not within the window range, update the tracking target to complete the continuous tracking of the tracking target.

[0006] Further, in Step 1, the feature extraction of all images includes: performing Gaussian filtering on all images to obtain the luminance and color-opponent dimension space, calculating the mean values of luminance, opponent dimensions and opponent dimensions in the luminance and color-opponent dimension space, and obtaining the feature values by combining the data fusion formula; The data fusion formula is expressed as: ; where, represents the feature value; represents the luminance, represents the mean value of the luminance; represents the opponent dimension , represents the opponent dimension and represents the mean value of the opponent dimension; represents the opponent dimension , represents the opponent dimension and represents the mean value of the opponent dimension.

[0007] Further, the determination of the image threshold based on the linear function of the standard deviation includes: using the initial kinetic energy in the electro-optical sensor electrical effect state of the single camera as the slope of the linear function, using the mean square deviation of the camera background noise of the single camera as the intercept of the linear function, and using the standard deviation as the input value of the linear function, and determining the output value of the linear function as the image threshold.

[0008] Further, the re-acquisition of the image captured by the optoelectronic detection device, separation of the moving points in combination with the image threshold, and recording of the deviation values include: re-acquiring the image captured by the optoelectronic detection device, constructing a normal distribution model based on the re-acquired image, calculating the variance value of the normal distribution model, separating the moving points exceeding the image threshold, and recording the deviation values.

[0009] Further, the windowing based on the coordinate position of the tracking target in combination with a preset distance includes: performing a windowing operation in the image based on the coordinate position of the tracking target in combination with a preset distance.

[0010] Further, step 4 includes: repeating steps 1 to 3 to obtain a new moving point, calculating whether the distance between the coordinates of the new moving point and the coordinates of the tracking target is greater than the preset distance of the windowing. If it is greater than the preset distance, update the new moving point as the tracking target, and re-window based on the preset distance to complete the continuous tracking of the tracking target.

[0011] Further, it also includes step 5: performing a third-order derivative calculation on the coordinate position of the tracking target to obtain the jerk, obtaining the aggregation degree based on the convolution calculation in step 3, and performing a category judgment on the tracking target based on the aggregation degree and the jerk.

[0012] Further, the category judgment of the tracking target based on the aggregation degree and the jerk includes: When the obtained aggregation degree varies within a plurality of repeated values of a predetermined number, it is determined to be a bird; When the jerk is lower than a predetermined threshold, it is determined to be a floating object; When the jerk is constant and the maintained direction does not change, it is determined to be a flight.

[0013] In a second aspect, the present invention also provides a staring-mode optoelectronic detection search and tracking calculation system, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the method described in any one of the above.

[0014] Beneficial effects: A staring-mode optoelectronic detection search and tracking calculation method and system provided by the present invention, through feature extraction, selects the deviation value of the moving point based on the feature, and then determines the tracking target among the deviation values of the moving points, effectively achieving the effect of a relatively static camera tracking a dynamic target, without the need for the camera to rotate at high speed following the tracking target, and solving the problem of blurred imaging in the existing optoelectronic detection search and tracking during the process of searching and tracking the target. Description of the drawings

[0015] Figure 1 It is a flowchart of a staring-mode optoelectronic detection search and tracking calculation method according to an embodiment of the present invention; Figure 2 It is a comparison schematic diagram before and after image feature extraction according to an embodiment of the present invention; Figure 3 It is a schematic diagram of constructing a normal distribution model for the variance value of the frame difference according to an embodiment of the present invention; Figure 4 It is a schematic diagram of separating moving points from an image threshold according to an embodiment of the present invention; In Figure 4 where A represents the separated moving point region, B represents the plane formed by the image threshold, and C represents the plane formed by the standard deviation of the newly obtained image; Figure 5 Schematic diagram for tracking different targets in the embodiments of the present invention; Figure 6 Schematic diagram of the motion blur state of traditional turntable search and tracking. Detailed implementation manners

[0016] The technical solutions of the present invention will be described clearly and completely below. Apparently, the described embodiments are only a part rather than all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without any creative efforts shall fall within the protection scope of the present invention.

[0017] Unless otherwise defined, the technical terms or scientific terms used in the present invention shall have the ordinary meanings understood by those of ordinary skill in the art to which the present invention belongs. The "first", "second" and similar terms used in the present invention do not denote any order, quantity or importance, but are only used to distinguish different components. Similarly, the terms such as "a" or "one" do not denote a quantity limitation, but mean that there is at least one. The terms such as "connected" or "coupled" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left" and "right" are only used to represent relative positional relationships, and when the absolute position of the object to be described changes, the relative positional relationships will also change accordingly.

[0018] Please refer to Figure 1 , an optoelectronic detection search and tracking calculation method in a staring mode provided by an embodiment of the present application is applied to an optoelectronic detection device composed of a single camera and an image processing host. In this embodiment, the image of the single camera has 5120 pixels in length and 5120 pixels in width, and the image with a depth of 12 bits per pixel. The 12-bit depth data sent by the camera is transmitted to the CPU of the image processing host.

[0019] The method includes the following steps: Step 1: Obtain a plurality of consecutive images captured by the optoelectronic detection device, extract features from all the images, and obtain the eigenvalue of the gradient change rule of all the images; Specifically, perform Gaussian filtering on all the images, obtain the luminance and color-opponent dimension space, calculate the means of the luminance, opponent dimension and opponent dimension in the luminance and color-opponent dimension space, and obtain the eigenvalue by combining the data fusion formula. Please refer to Figure 2 , in Figure 2 the left figure is the original image, and the right figure is the result after feature extraction; The data fusion formula is expressed as: ; Among them, represents the eigenvalue; represents the brightness, represents the mean value of the brightness; represents the opposition dimension , represents the opposition dimension of the mean value; represents the opposition dimension , represents the opposition dimension of the mean value; Step 2: Calculate the frame difference for the eigenvalues of all adjacent images and construct a normal distribution model, calculate the standard deviation of the variance value of the normal distribution model, determine the image threshold based on the linear function of the standard deviation, re-obtain the images captured by the optoelectronic detection device, and separate the moving points in combination with the image threshold and record the deviation value; Please refer to Figure 3 , calculate the frame difference for the eigenvalues of all adjacent images and construct a normal distribution model, and calculate the standard deviation of the variance value of the normal distribution model; For the linear function of the standard deviation, use the initial kinetic energy in the electro - effect state of the optoelectronic sensor of a single camera as the slope of the linear function, use the mean square deviation of the camera background noise of a single camera as the intercept of the linear function, and use the standard deviation as the input value of the linear function , and the output value of the linear function is determined as the image threshold; Please refer to Figure 4 , after determining the image threshold based on the linear function of the standard deviation, re - obtain the images captured by the optoelectronic detection device, and construct a normal distribution model based on the re - obtained images, calculate the variance value of the normal distribution model, separate the moving points exceeding the image threshold and record the deviation value; Step 3: Perform convolution calculation on the deviation values of all moving points to obtain the deviation degree, and select the moving point corresponding to the maximum deviation degree as the tracking target; Perform convolution calculation based on the principle of aggregate probability statistics. When the probability of a point becoming a suspicious target is a fixed value, then the probability that multiple points in the 3 * 3 square during convolution calculation are all suspicious targets in the same frame is infinitely close to 100%.

[0020] Step 4: Based on the coordinate position of the tracking target, perform windowing in combination with a preset distance, then repeat Steps 1 to 3 to obtain new moving points and determine whether they are within the window range. If not within the window range, update the tracking target to complete the continuous tracking of the tracking target.

[0021] Specifically, repeat steps 1 to 3 to obtain a new moving point, and calculate whether the distance between the coordinates of the new moving point and the coordinates of the tracking target is greater than the preset distance for windowing. If it is greater than the preset distance, update the new moving point as the tracking target, and re-window in combination with the preset distance to complete the continuous tracking of the tracking target.

[0022] Step 5: Perform a third-order derivative calculation on the coordinate position of the tracking target to obtain the jerk, obtain the aggregation degree based on the convolution calculation in step 3, and perform a category judgment on the tracking target based on the aggregation degree and the jerk.

[0023] Specifically, if the obtained aggregation degree varies within a predetermined number of repeated values, it is determined to be a bird; if the jerk is lower than a predetermined threshold, it is determined to be a floating object; if the jerk is constant and the direction remains unchanged, it is determined to be a flight.

[0024] For the case where the jerk is lower than a predetermined threshold and is determined to be a floating object, other corresponding thresholds can also be set. When the jerk is higher than this threshold, it can be determined as a false alarm and the tracking is stopped.

[0025] Please refer to Figure 5 , the large left interface area shows the entire camera image reduced by 64 times, and the upper and lower small areas on the right show the local original size images of the target tracking area. The gaze search and tracking system continuously detects moving targets within the entire field of view and continuously tracks the drones within the field of view. Even if the flight direction of the bird is different from that of the drone, it can also search, detect, and continuously track. The stationary building in the lower left corner does not affect the continuous search and tracking of the drones and birds by the system.

[0026] Please refer to Figure 6 , the motion blur schematic diagram of the traditional turntable. In the left figure, the falcon being tracked by the turntable flies over the building at high speed. The falcon with the same angular velocity relative to the turntable is clear, but the building behind the falcon is significantly blurred with trailing and tilted in the direction of motion. In the right figure, the motion direction of the falcon is inconsistent with that of the camera, and the camera quickly zooms in on the direction of the falcon, so the falcon is blurred. Such blurring is likely to cause the failure of this tracking. The method to prevent motion blur is to reduce the camera exposure time and reduce relative motion. Reducing the camera resolution is an effective method to reduce the camera exposure time, and holding one's breath during shooting is also an effective method to prevent motion blur. The example of this application uses the gaze method to eliminate the motion blur of camera imaging and greatly improves the performance of search and tracking.

[0027] The preferred specific embodiments of the present invention have been described in detail above. It should be understood that those of ordinary skill in the art can make many modifications and variations based on the concept of the present invention without creative efforts. Therefore, all technical solutions that can be obtained by those skilled in the art in the technical field according to the concept of the present invention through logical analysis, reasoning, or limited experiments on the basis of the prior art shall fall within the protection scope determined by the claims.

Claims

1. A gaze-based photoelectric detection search and tracking calculation method, characterized in that: The photoelectric detection device, which is applied to a single camera and an image processing host, includes the following steps: Step 1: Obtain several continuous images taken by the photoelectric detection device, perform feature extraction on all images, and obtain the characteristic values ​​of the gradient change law of all images; Step 2: Calculate the frame difference of the feature values ​​of all adjacent images and construct a normal distribution model, calculate the standard deviation of the variance value of the normal distribution model, determine the image threshold based on the linear function of the standard deviation, reacquire the image taken by the photoelectric detection device, combine the image threshold to separate the moving points and record the deviation value; Step 3: Perform convolution calculation on the deviation values ​​of all moving points to obtain the deviation, and select the moving point corresponding to the maximum deviation as the tracking target; Step 4: Open a window based on the coordinate position of the tracking target and the preset distance, and repeat steps 1 to 3 to obtain a new moving point and determine whether it is within the window range. If it is not within the window range, update the tracking target to complete the continuous tracking of the tracking target.

2. The gaze-mode photoelectric detection search and tracking calculation method according to claim 1, characterized in that: In step 1, the feature extraction of all images includes: performing Gaussian filtering on all images, obtaining brightness and color-opposition dimension space, calculating brightness and color-opposition dimension in the brightness and color-opposition dimension space And the opposite dimension The mean of is obtained by combining the data fusion formula to obtain the eigenvalue; The data fusion formula is expressed as: ； in, represents the eigenvalue; Indicates brightness, represents the mean value of brightness; Representing opposing dimensions , Representing opposing dimensions The mean of Representing opposing dimensions , Representing opposing dimensions The mean of .

3. The gaze-mode photoelectric detection search and tracking calculation method according to claim 1, characterized in that: The method of determining the image threshold value based on the linear function of the standard deviation includes: taking the initial kinetic energy of the photoelectric sensor of the single camera in the electric effect state as the slope of the linear function, taking the mean square error of the camera background noise of the single camera as the intercept of the linear function, and taking the standard deviation as the input value of the linear function. , the output value of a function Determine the image threshold.

4. The gaze-mode photoelectric detection search and tracking calculation method according to claim 3 is characterized in that: The method of reacquiring an image taken by a photoelectric detection device and combining it with an image threshold to separate moving points and record deviation values ​​includes: reacquiring an image taken by a photoelectric detection device, constructing a normal distribution model based on the reacquired image, calculating the variance value of the normal distribution model, separating moving points that exceed the image threshold and recording the deviation value.

5. The gaze-mode photoelectric detection search and tracking calculation method according to any one of claims 1 to 4, characterized in that: The performing windowing based on the coordinate position of the tracking target in combination with the preset distance includes: performing a windowing operation in the image based on the coordinate position of the tracking target in combination with the preset distance.

6. The gaze-mode photoelectric detection search and tracking calculation method according to claim 5, characterized in that: The step 4 includes: repeating steps 1 to 3 to obtain a new moving point, calculating whether the distance between the coordinates of the new moving point and the coordinates of the tracking target is greater than the preset distance of the window, and if it is greater than the preset distance, updating the new moving point as the tracking target, and re-opening the window in combination with the preset distance to complete the continuous tracking of the tracking target.

7. The gaze-mode photoelectric detection search and tracking calculation method according to claim 1, characterized in that: The method further includes step 5: performing a third-order derivative calculation on the coordinate position of the tracked target to obtain the jerk, obtaining the degree of aggregation based on the convolution calculation in step 3, and performing a category judgment on the tracked target based on the degree of aggregation and the jerk.

8. The gaze-mode photoelectric detection search and tracking calculation method according to claim 7, characterized in that: The aggregation degree and the jerk determine the category of the tracked target, including: When the obtained degree of polymerization varies within a predetermined number of repeated values, it is determined to be a bird; When the jerk is lower than a predetermined threshold, it is determined to be a floating object; When the jerk is constant and the direction does not change, it is judged as a flight.

9. A gaze-mode photoelectric detection search and tracking computing system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method described in any one of claims 1 to 8 are implemented.

Images