Video jitter fast detection method, device, computer equipment and medium

Abstract

The present application provides a method for rapid detection of video jitter. The method intercepts segmented videos in surveillance videos in real time, obtains frame images of the segmented videos, then performs feature point matching on frame images of two adjacent frames in the segmented videos, then calculates the number of mismatched point pairs of each frame image and obtains the distribution of mismatched point pairs of the segmented videos, calculates the image quality of each frame image and obtains the image quality distribution of the segmented videos, and finally generates a video jitter recognition signal if the distribution of mismatched point pairs is consistent with the image quality distribution. Whether the video is jittery is determined by the distribution of mismatched point pairs and the image quality distribution, thereby avoiding the problem of inaccurate results caused by feature point selection and object movement in the picture, and improving the accuracy of detection.

Description

Technical Field

[0001] The present invention belongs to the technical field of video jitter, and in particular relates to a method, device, computer equipment and medium for rapid detection of video jitter. Background Art

[0002] As the demand for video surveillance continues to increase, existing video surveillance needs to be maintained in a timely manner. Due to the huge demand for video surveillance, the maintenance workload also increases when faced with maintenance, making the maintenance efficiency of the video surveillance system an issue that needs to be urgently addressed.

[0003] In the process of video surveillance, video surveillance anomalies are the most common phenomenon, and video surveillance anomalies include: video surveillance screen anomalies, video surveillance occlusion anomalies, video surveillance jitter, etc. Among them, the causes of video surveillance jitter usually exist in the following two categories:

[0004] 1. During the installation process, the video surveillance equipment was not installed tightly, resulting in looseness between the video surveillance equipment and the mounting bracket, causing the video image to shake during the shooting process.

[0005] 2. During use, due to vibrations in the surrounding environment, the video surveillance equipment will produce video shaking during shooting.

[0006] The existing technology usually sets feature points for adjacent frames of video images, and identifies whether jitter occurs by detecting the displacement distance between adjacent feature points. However, this method will have errors because it only compares and judges by the displacement distance between feature points. That is, the jitter of the frame image will lead to inaccurate selection of feature points, which will lead to inaccurate judgment results; secondly, only using feature points for comparison will result in a single reference basis for comparison, resulting in inaccurate comparison results; there are errors in the selection of feature points and the calculation of the displacement of feature points of adjacent frames, resulting in the comparison process. Due to the small calculation difference of the displacement of feature points of adjacent frames, the error range is further increased, which in turn causes differences in the judgment of the results; secondly, in the frame image of video surveillance, the movement of objects (such as pedestrians, vehicles, etc.) will affect the selection and displacement of feature points, resulting in misjudgment of the subsequent judgment of whether the video surveillance is jittery. Summary of the invention

[0007] In view of the deficiencies of the above-mentioned prior art, the purpose of the present invention is to provide a method, device, computer equipment and medium for rapid detection of video jitter, so as to solve the problem that the selection and displacement of feature points are affected by the single reference basis of feature points and the movement of objects (such as pedestrians, vehicles, etc.), thereby causing misjudgment of the result of whether the video surveillance is jittery.

[0008] The present invention provides a method for rapid detection of video jitter, comprising real-time interception of segmented videos in a surveillance video, and obtaining frame images of the segmented videos;

[0009] Perform feature point matching on frame images of two adjacent frames in the segmented video;

[0010] Calculate the number of mismatched point pairs in each frame image and obtain the distribution of mismatched point pairs in the segmented video;

[0011] Calculate the image quality of each frame and obtain the image quality distribution of the segmented video;

[0012] If the distribution of mismatched point pairs is consistent with the image quality distribution, a video jitter recognition signal is generated.

[0013] In one embodiment of the present invention, the step of capturing the segmented video in the surveillance video in real time includes:

[0014] The unit time of the segmented video is obtained, and a plurality of segmented videos having a duration equal to the unit time are obtained in sequence and continuously.

[0015] In one embodiment of the present invention, the step of matching feature points of two adjacent frames in the segmented video includes:

[0016] ORB feature point detection and brute force matching are used for the frame images of two adjacent frames in the segmented video.

[0017] In one embodiment of the present invention, the step of calculating the number of mismatched point pairs of each frame image and obtaining the distribution of mismatched point pairs of the segmented video includes:

[0018] Obtain a horizontally stitched image of two adjacent frame images;

[0019] Establish the coordinate system of the horizontally stitched image and the coordinate data of the matching point pairs;

[0020] Get the slope set of matching point pairs;

[0021] Get the mismatch slope difference threshold;

[0022] Obtain the number of mismatched point pairs of the frame image of the current frame according to the slope set and the mismatched slope difference threshold;

[0023] Get the number of mismatched point pairs of all frame images of the segmented video;

[0024] Get the distribution of mismatched point pairs for segmented videos.

[0025] In one embodiment of the present invention, the step of calculating the image quality of each frame image and obtaining the image quality distribution of the segmented video includes:

[0026] The energy gradient function value of the frame image of each frame is obtained, and the image quality distribution is obtained according to the energy gradient function value.

[0027] In one embodiment of the present invention, the step of obtaining the energy gradient function value of each frame image, and obtaining the image quality distribution according to the energy gradient function value comprises:

[0028] Obtaining the grayscale value of the frame image of the previous frame and the grayscale value of the frame image of the next frame between two adjacent frames;

[0029] Obtaining grayscale gradient values ​​of two adjacent frame images according to grayscale values ​​of the frame images of two adjacent frames, and using the grayscale gradient values ​​as energy gradient function values ​​of the two adjacent frame images;

[0030] The image quality distribution is obtained according to the energy gradient function value.

[0031] In one embodiment of the present invention, the step of generating a video jitter identification signal if the distribution of mismatched point pairs is consistent with the image quality distribution includes:

[0032] The data stability of the distribution of mismatched point pairs and image quality distribution is checked. If the data stability exceeds the detection threshold, a video jitter recognition signal is generated.

[0033] The present invention provides a video jitter rapid detection device and a video capture module for capturing segmented videos in a surveillance video;

[0034] A feature point matching module is used to match feature points of two adjacent frames in the segmented video;

[0035] The calculation module is used to calculate the number of mismatched point pairs in each frame image and obtain the distribution of mismatched point pairs in the segmented video and the image quality of each frame image and obtain the image quality distribution of the segmented video.

[0036] The result judgment module is used to judge the consistency between the distribution of mismatched point pairs and the image quality distribution and generate a video jitter recognition signal.

[0037] The present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the method described above.

[0038] The present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the method described above.

[0039] Implementing the embodiments of the present invention will have at least the following beneficial effects:

[0040] The present invention provides a video jitter fast detection method. The method intercepts segmented videos in monitoring videos in real time to obtain frame images of the segmented videos, then performs feature point matching on frame images of two adjacent frames in the segmented videos, then calculates the number of mismatched point pairs of each frame image and obtains the mismatched point pair distribution of the segmented videos, calculates the image quality of each frame image and obtains the image quality distribution of the segmented videos, and finally generates a video jitter recognition signal if the mismatched point pair distribution is consistent with the image quality distribution, and judges whether the video is jittering by the mismatched point pair distribution and the image quality distribution, thereby avoiding the problem of inaccurate results caused by feature point selection and object movement in the picture, and improving the accuracy of detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0042] Figure 1 It is a flowchart of an embodiment of a method for rapid video jitter detection in the present invention;

[0043] Figure 2 It is a flowchart of another embodiment of the method for rapid detection of video jitter in the present invention;

[0044] Figure 3 It is a structural block diagram of the video jitter fast detection device in the present invention;

[0045] Figure 4 It is a structural block diagram of the computer device in the present invention.

[0046] In the figure:

[0047] Video jitter fast detection device 100, video capture module 110, feature point matching module 120, calculation module 130, result judgment module 140, computer device 200, memory 210, processor 220. DETAILED DESCRIPTION

[0048] In order to enable those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of this application.

[0049] It should be noted that when an element is referred to as being "fixed on" or "set on" another component, it can be directly on the other component or indirectly set on the other component; when a component is referred to as being "connected to" another component, it can be directly connected to the other component or indirectly connected to the other component.

[0050] It should be understood that the terms "width", "height", "up", "down", "front", "back", "left", "right", "vertical", "

[0051] The orientations or positional relationships indicated by “horizontal”, “top”, “bottom”, “inside”, “outside”, etc. are based on the orientations or positional relationships shown in the drawings and are only for the convenience of describing the present application and simplifying the description. They do not indicate or imply that the devices or components referred to must have a specific orientation, be constructed and operate in a specific orientation. Therefore, they should not be understood as limitations on the present application.

[0052] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "multiple" and "several" mean two or more, unless otherwise clearly and specifically defined.

[0053] It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the conditions under which this application can be implemented. Therefore, they have no substantive technical significance. Any structural modification, change in proportional relationship or adjustment of size should still fall within the scope of the technical content disclosed in this application without affecting the effects and purposes that can be achieved by this application.

[0054] The present invention provides a method for quickly detecting video jitter, comprising:

[0055] S101, intercepting segmented videos in the surveillance video in real time to obtain frame images of the segmented videos;

[0056] S102, performing feature point matching on frame images of two adjacent frames in the segmented video;

[0057] S103, calculating the number of mismatched point pairs in each frame image and obtaining the distribution of mismatched point pairs in the segmented video;

[0058] S104, calculating the image quality of each frame image and obtaining the image quality distribution of the segmented video;

[0059] S105: If the distribution of the mismatched point pairs is consistent with the image quality distribution, a video jitter recognition signal is generated.

[0060] In one embodiment of the present invention, the step of capturing the segmented video in the surveillance video in real time includes:

[0061] The unit time of the segmented video is obtained, and a plurality of segmented videos having a duration equal to the unit time are obtained in sequence and continuously.

[0062] Embodiment 1:

[0063] In this embodiment, the method captures segmented videos in surveillance videos in real time, obtains frame images of the segmented videos, then performs feature point matching on frame images of two adjacent frames in the segmented videos, then calculates the number of mismatched point pairs of each frame image and obtains the mismatched point pair distribution of the segmented videos, calculates the image quality of each frame image and obtains the image quality distribution of the segmented videos, and finally generates a video jitter recognition signal if the mismatched point pair distribution is consistent with the image quality distribution, and determines whether the video is jittering based on the mismatched point pair distribution and the image quality distribution, thereby avoiding the problem of inaccurate results caused by feature point selection and object movement in the picture, and improving the accuracy of detection.

[0064] Specifically, the real-time capture of segmented videos in surveillance videos is to read the video surveillance device in real time or capture any video segment from the recording of the video surveillance device, and the captured video or the read video is the segmented video; then the segmented video is deframed to obtain the frame image of the segmented video.

[0065] In one embodiment of the present invention, the step of matching feature points of two adjacent frames in the segmented video includes:

[0066] ORB feature point detection and brute force matching are used for the frame images of two adjacent frames in the segmented video.

[0067] Specifically, ORB feature point detection and brute force matching are performed on frame images of two adjacent frames in the segmented video, so as to perform feature point matching on two adjacent frame images in the segmented video.

[0068] In one embodiment of the present invention, the step of calculating the number of mismatched point pairs of each frame image and obtaining the distribution of mismatched point pairs of the segmented video includes:

[0069] Obtain a horizontally stitched image of two adjacent frame images;

[0070] Establish the coordinate system of the horizontally stitched image and the coordinate data of the matching point pairs;

[0071] Get the slope set of matching point pairs;

[0072] Get the mismatch slope difference threshold;

[0073] Obtain the number of mismatched point pairs of the frame image of the current frame according to the slope set and the mismatched slope difference threshold;

[0074] Get the number of mismatched point pairs of all frame images of the segmented video;

[0075] Get the distribution of mismatched point pairs for segmented videos.

[0076] Specifically, the real-time capture of segmented video in the surveillance video is to capture any video segment by real-time reading of the video surveillance device or from the recording of the video surveillance device, and the captured video or the read video is the segmented video; then the segmented video is deframed to obtain the frame image of the segmented video, and then the frame images of two adjacent frames in the segmented video are matched with feature points, that is, the frame images of each of the two adjacent frames are horizontally spliced, a coordinate system is established in the spliced ​​image, the upper left corner of the spliced ​​image is used as the coordinate origin, the width of the spliced ​​image is used as the X-axis, and the height of the spliced ​​image is used as the Y-axis, the coordinate data of the matching point pairs in the coordinate system are obtained, and the slope of the line connecting each matching point pair is calculated to obtain the matching point pair. Slope set, statistics are performed on the slope set to obtain the frequency of each slope, and the one with the highest frequency is selected as the reference slope. The difference between the reference slope and each slope in the slope set is calculated, and the matching point pairs corresponding to the slopes whose difference is greater than the mismatch slope difference threshold are recorded as mismatch point pairs. The frequencies corresponding to all mismatch slopes are accumulated to obtain the number of mismatch point pairs, and the calculation is performed in sequence. The number of mismatch point pairs calculated by matching the feature points of the current frame image with the feature points of the next frame is taken to obtain the number of mismatch point pairs of the frame image of the current frame, and the number of mismatch point pairs of all frame images of the segmented video is calculated in sequence to obtain the number of mismatch point pairs of all frame images of the segmented video, and the distribution of mismatch point pairs of the segmented video is obtained.

[0077] It should be noted that the server adopts the method of horizontally stitching images to facilitate the establishment of a coordinate system for horizontally stitched images, and then obtains the coordinate data of the matching point pairs located in the coordinate system, so that the straight line between the two coordinate points is obtained through the coordinates of the matching point pairs, and then the slope of the straight line between the two coordinate points is obtained. The slope of each matching point pair is calculated to obtain the slope set of the matching point pairs. The mismatch slope difference threshold is a manually preset difference. According to the slope set and the mismatch slope difference threshold, the number of mismatch point pairs in the frame image of the current frame can be screened out, and the distribution of mismatch point pairs in the segmented video can be obtained based on all the mismatch point pairs.

[0078] In one embodiment of the present invention, the step of calculating the image quality of each frame image and obtaining the image quality distribution of the segmented video includes:

[0079] The energy gradient function value of the frame image of each frame is obtained, and the image quality distribution is obtained according to the energy gradient function value.

[0080] Specifically, the server obtains the energy gradient function value of the frame image of each frame, obtains the image quality distribution information according to the energy gradient function value, and then judges the image quality distribution information to obtain the stability of the image quality distribution in the segmented video image.

[0081] In one embodiment of the present invention, the step of obtaining the energy gradient function value of each frame image, and obtaining the image quality distribution according to the energy gradient function value comprises:

[0082] Obtaining the grayscale value of the frame image of the previous frame and the grayscale value of the frame image of the next frame between two adjacent frames;

[0083] Obtaining grayscale gradient values ​​of two adjacent frame images according to grayscale values ​​of the frame images of two adjacent frames, and using the grayscale gradient values ​​as energy gradient function values ​​of the two adjacent frame images;

[0084] The image quality distribution is obtained according to the energy gradient function value.

[0085] Specifically, the grayscale value of the frame image of the previous frame of two adjacent frames is obtained, and the grayscale value of the frame image of the next frame is obtained, that is, the quality of the image is quantified by the energy gradient function value, the image is regarded as a two-dimensional discrete matrix, the grayscale information of the frame image is obtained by using the gradient function, the width of the frame image is used as the X direction, the height of the frame image is used as the Y direction, the difference between the grayscale values ​​of adjacent pixels in the X direction and the Y direction is calculated to obtain the square sum result, the calculated square sum result is used as the gradient value of each pixel point, the grayscale gradient value of the two adjacent frames is obtained according to the grayscale values ​​of the frame images of the two adjacent frames, the grayscale gradient value is used as the energy gradient function value of the two adjacent frames, and the image quality distribution is obtained according to the energy gradient function value.

[0086] It should be noted that when there is no jitter, the clarity of multiple consecutive frames of images is not much different, but when jitter occurs, the image may be blurred, which in turn causes large fluctuations between adjacent frames. The edge of the clear image in focus is sharper and clearer than that of the defocused image in the blurred state, that is, the grayscale value changes greatly, so there is a larger gradient value.

[0087] In one embodiment of the present invention, the step of generating a video jitter identification signal if the distribution of mismatched point pairs is consistent with the image quality distribution includes:

[0088] The data stability of the distribution of mismatched point pairs and image quality distribution is checked. If the data stability exceeds the detection threshold, a video jitter recognition signal is generated.

[0089] Embodiment 2:

[0090] In this embodiment, the real-time capture of the segmented video in the surveillance video is to capture any video segment by real-time reading of the video surveillance device or from the recording of the video surveillance device, and the captured video or the read video is the segmented video; then the segmented video is deframed to obtain the frame image of the segmented video, and then the frame images of two adjacent frames in the segmented video are matched with feature points, that is, the frame images of each two adjacent frames are horizontally spliced, a coordinate system is established in the spliced ​​image, the upper left corner of the spliced ​​image is used as the coordinate origin, the width of the spliced ​​image is used as the X-axis, and the height of the spliced ​​image is used as the Y-axis, the coordinate data of the matching point pairs in the coordinate system are obtained, and the slope of the line connecting each matching point pair is calculated to obtain the matching point. For a slope set, statistics are performed on the slope set to obtain the frequencies corresponding to each slope, and the one with the highest frequency is selected as the reference slope. The reference slope is calculated to differ from each slope in the slope set, and the matching point pairs corresponding to the slopes whose differences are greater than the mismatch slope difference threshold are recorded as mismatch point pairs. The frequencies corresponding to all mismatch slopes are accumulated to obtain the number of mismatch point pairs, and the number of mismatch point pairs calculated by matching the feature points of the current frame image with the feature points of the next frame is calculated in sequence, so as to obtain the number of mismatch point pairs of the frame image of the current frame, and the number of mismatch point pairs of all frame images of the segmented video is calculated in sequence, so as to obtain the number of mismatch point pairs of all frame images of the segmented video, and obtain the distribution of mismatch point pairs of the segmented video;

[0091] Obtain the grayscale value of the frame image of the previous frame of two adjacent frames, and obtain the grayscale value of the frame image of the next frame, that is, use the energy gradient function value to quantify the quality of the image, regard the image as a two-dimensional discrete matrix, use the gradient function to obtain the grayscale information of the frame image, use the width of the frame image as the X direction, and the height of the frame image as the Y direction, calculate the difference between the grayscale values ​​of adjacent pixels in the X direction and the Y direction to obtain the square sum result, use the calculated square sum result as the gradient value of each pixel point, obtain the grayscale gradient value of the two adjacent frame images according to the grayscale values ​​of the two adjacent frame images, use the grayscale gradient value as the energy gradient function value of the two adjacent frame images, and obtain the image quality distribution according to the energy gradient function value;

[0092] Check the data stationarity of the distribution of mismatched point pairs and image quality distribution, that is, if the data stationarity of the distribution of mismatched point pairs and image quality distribution exceeds the detection threshold, a video jitter recognition signal is generated, and it is judged that the video surveillance equipment is jittering, and an alarm signal is issued in time, that is, the number of mismatched point pairs is calculated for two adjacent frames, so the number of mismatched point pairs in N frames is N-1. When verifying the data stationarity of the distribution of mismatched point pairs and image quality distribution, when the test result of the data stationarity satisfies the pvalue value not less than the threshold value or the adf value not less than 1% of the statistical value, the data is judged to be unstable. If the data stationarity of the distribution of mismatched point pairs and the image quality distribution are both unstable, it indicates that the video is jittering.

[0093] Specifically, the data stability of the distribution of mismatched point pairs and the image quality distribution is checked, and if the data stability exceeds a detection threshold, a video jitter recognition signal is generated.

[0094] At least two methods are included in obtaining the distribution of mismatched point pairs and the image quality distribution, and performing statistical analysis on the consistency between the distribution of mismatched point pairs and the image quality distribution.

[0095] In one embodiment, the server counts the number of mismatched point pairs of each frame image in the segmented video in real time, and generates a distribution of the number of point pairs according to the frame sequence based on the number of mismatched point pairs in the segmented video. The distribution of the number of point pairs can be a point pair distribution curve.

[0096] The server simultaneously counts the image quality gradient data of each frame image in the segmented video, and generates the distribution of gradient data based on the image quality gradient data in the segmented video according to the frame sequence, wherein the distribution of the gradient data can be a gradient data distribution curve;

[0097] A unit time period of the segmented video is set. If the point pair distribution curve continuously exceeds the point pair threshold and the gradient data distribution curve does not continuously exceed the gradient data threshold within the unit time period, the distribution of the mismatched point pairs is determined to be consistent with the image quality distribution, and a video jitter recognition signal is generated. That is, when jittering, the image quality will be lower than the normal image quality, and the distribution curve will also show a downward fluctuation trend.

[0098] By counting the number of mismatched point pairs within a unit time period that continuously exceeds the point pair threshold, the presence of moving objects in the image can be detected. However, the blurred frame image due to movement cannot be very accurately determined whether it is partial movement or lens jitter. Therefore, the gradient data distribution is further used to further increase the basis for determination. By evaluating the overall image quality of the frame image, the lens jitter misjudgment caused by the increase in the number of mismatched point pairs due to the movement of partial objects can be screened out. Therefore, the above scheme can enhance the accuracy of video jitter recognition.

[0099] In another embodiment, the server counts the number of mismatched point pairs of each frame image in the segmented video in real time, generates a distribution of the number of mismatched point pairs in the segmented video according to the frame sequence, and calculates the difference in coordinates of mismatched point pairs of two adjacent frames of the frame image to obtain the difference information of mismatched point pairs to be screened;

[0100] The server simultaneously counts the image quality gradient data of each frame in the segmented video, generates the distribution of gradient data based on the frame sequence of the image quality gradient data in the segmented video, and calculates the difference between the frame image quality gradient data of two adjacent frames to obtain the gradient data difference information to be screened;

[0101] Sorting the mismatched point pair difference information to be screened according to the frame image shooting time to generate sorted mismatched point pair difference information;

[0102] Sorting the gradient data difference information to be screened according to the frame image shooting time to generate sorted gradient data difference information;

[0103] Generate the mismatched point pair data difference variation according to the sorted mismatched point pair difference information;

[0104] Generating a gradient data difference change amount according to the sorted gradient data difference information;

[0105] The server analyzes the difference information of the sorted mismatched point pairs. If the difference information of the sorted mismatched point pairs shows that the difference of the mismatched point pairs increases continuously, and the difference of the mismatched point pairs shows a continuous decrease after the difference of the mismatched point pairs increases, the mismatched point pairs are judged to be unstable. The server analyzes the gradient data difference information after sorting. If the difference of the gradient data increases continuously after sorting, and the difference of the gradient data shows a continuous decrease, the image quality is judged to be unstable. If the mismatched point pairs and the image quality are both unstable, it is judged to be jittery of the video surveillance device. That is, in the process of lens jitter, the lens will first produce a larger jitter, and then produce a smaller jitter, so that the image quality will have a process of decline, and the distribution of the mismatched point pairs will have an obvious fluctuation.

[0106] By counting the number of mismatched point pairs within a unit time period that continuously exceeds the point pair threshold, the presence of a moving object in the image can be detected. However, because there is a moving object in the image, the moving object may have accelerated motion, decelerated motion, or uniform motion in the image. At this time, the feature points in the image will also be affected by the state of the moving object. At this time, the video surveillance device is not shaking. If the jitter is analyzed only by feature points, then when the moving object has accelerated motion, decelerated motion, or uniform motion in the image, the feature points will be misjudged due to the acceleration and deceleration of the moving object in the image, and thus misjudged as a jitter signal, resulting in inaccurate jitter judgment. The blurred frame image due to movement cannot be accurately judged as partial movement or lens shaking, so the data difference change of the mismatched point pair and the data difference change of the gradient data are further judged. When the video shakes, an acceleration and a deceleration shake will be generated. When the shake occurs, the feature points in the segmented image will have a relatively large displacement. In the process of shaking, the displacement trajectory of the feature points in the segmented image will obviously produce a back and forth shake relative to the static state coordinates, so that the displacement difference is very large, causing the data difference change of the mismatched point pair to first produce a displacement of a larger increment or a decrement relative to the static state, and then produce an opposite increment or a decrement. At the same time, the image quality will also produce a larger increment relative to the static state during this period of time, and then produce a decrement, resulting in the mismatched point pair and the gradient data to increase the increment and then decrease the change to increase the judgment basis, which can screen out the lens shake misjudgment caused by the increase in the number of matching point pairs due to the rapid movement of some objects. Therefore, the above scheme can further enhance the accuracy of video shake recognition.

[0107] The server generates a video jitter identification signal based on the judgment result to prompt the staff to carry out maintenance.

[0108] The present invention provides a video jitter fast detection device 100, a video capture module 110, which is used to capture segmented videos in a surveillance video;

[0109] A feature point matching module 120 is used to perform feature point matching on frame images of two adjacent frames in the segmented video;

[0110] A calculation module 130 is used to calculate the number of mismatched point pairs of each frame image and obtain the distribution of mismatched point pairs of the segmented video and the image quality of each frame image and obtain the image quality distribution of the segmented video;

[0111] The result judgment module 140 is used to judge the consistency between the distribution of mismatched point pairs and the image quality distribution and generate a video jitter recognition signal.

[0112] Embodiment three:

[0113] In this embodiment, the video jitter rapid detection device 100 captures the segmented video in the surveillance video through the video capture module 110, the feature point matching module 120 is used to perform feature point matching on the frame images of two adjacent frames in the segmented video, the calculation module 130 is used to calculate the number of mismatched point pairs of each frame image and obtain the mismatched point pair distribution of the segmented video and the image quality of each frame image and obtain the image quality distribution of the segmented video, the result judgment module 140 judges the consistency based on the mismatched point pair distribution and the image quality distribution and generates a video jitter recognition signal, thereby avoiding the problem of inaccurate results caused by feature point selection and object movement in the picture, and improving the accuracy of detection.

[0114] The present invention provides a computer device 200, including a memory 210 and a processor 220, wherein the memory 210 stores a computer program, and when the computer program is executed by the processor 220, the processor executes the method described above.

[0115] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment system can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory 220, storage, database or other media used in the embodiments provided in the present application may include non-volatile and / or volatile memory 220. The non-volatile memory 220 may include a read-only memory 220 (ROM), a programmable ROM (PROM), an electrically programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM) or a flash memory. The volatile memory 220 may include a random access memory 220 (RAM) or an external cache memory 220. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory 220 bus (Rambus) direct RAM (RDRAM), direct memory 220 bus dynamic RAM (DRDRAM), and memory 220 bus dynamic RAM (RDRAM), etc.

[0116] The present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the method described above.

[0117] Computer devices can be computing devices such as desktop computers, notebooks, PDAs, and cloud servers. Computer devices can interact with users through keyboards, mice, remote controls, touch pads, or voice control devices.

[0118] The computer device includes a processor, a memory and a network interface connected via a terminal bus. The memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device has a storage operation terminal and may also store a computer program. When the computer program is executed by the processor, the processor can implement the above-mentioned video jitter fast detection method. The internal memory may also store a computer program. When the computer program is executed by the processor, the processor can implement the above-mentioned video jitter fast detection method. Those skilled in the art can understand that Figure 4 The structure shown in the figure is merely a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the device to which the scheme of the present application is applied. The specific device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0119] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for rapid detection of video jitter, characterized in that: include: Intercept the segmented video in the surveillance video in real time and obtain the frame image of the segmented video; Perform feature point matching on frame images of two adjacent frames in the segmented video; Calculate the number of mismatched point pairs in each frame image and obtain the distribution of mismatched point pairs in the segmented video; Calculate the image quality of each frame and obtain the image quality distribution of the segmented video; If the distribution of mismatched point pairs is consistent with the image quality distribution, a video jitter recognition signal is generated; Wherein, the step of generating a video jitter identification signal if the distribution of mismatched point pairs is consistent with the image quality distribution comprises: The server counts the number of mismatched point pairs of each frame image in the segmented video in real time, and generates a distribution of point pairs based on the number of mismatched point pairs in the segmented video according to the frame sequence, and the distribution of the point pairs is a point pair distribution curve; The server simultaneously counts the image quality gradient data of each frame image in the segmented video, and generates the distribution of gradient data based on the image quality gradient data in the segmented video according to the frame sequence, wherein the distribution of the gradient data is a gradient data distribution curve; A unit time period of the segmented video is set. If the point pair distribution curve continuously exceeds the point pair threshold and the gradient data distribution curve does not continuously exceed the gradient data threshold within the unit time period, the distribution of the mismatched point pairs is determined to be consistent with the image quality distribution, and a video jitter recognition signal is generated.

2. The video jitter fast detection method according to claim 1, characterized in that: The step of intercepting the segmented video in the surveillance video in real time comprises: The unit time of the segmented video is obtained, and multiple segmented videos with a duration of the unit time are obtained in sequence and continuously.

3. The method for rapid video jitter detection according to claim 1, characterized in that: The step of matching feature points of two adjacent frames in the segmented video includes: ORB feature point detection and brute force matching are used for the frame images of two adjacent frames in the segmented video.

4. The method for rapid video jitter detection according to claim 1, characterized in that: The step of calculating the number of mismatched point pairs of each frame image and obtaining the distribution of mismatched point pairs of the segmented video includes: Obtain a horizontally stitched image of two adjacent frame images; Establish the coordinate system of the horizontally stitched image and the coordinate data of the matching point pairs; Get the slope set of matching point pairs; Get the mismatch slope difference threshold; Obtain the number of mismatched point pairs of the frame image of the current frame according to the slope set and the mismatched slope difference threshold; Get the number of mismatched point pairs of all frame images of the segmented video; Get the distribution of mismatched point pairs for segmented videos.

5. A video jitter rapid detection device, characterized in that: The method for quickly detecting video jitter according to any one of claims 1 to 4 is used, wherein the device for quickly detecting video jitter comprises: A video capture module is used to capture segmented videos in surveillance videos; A feature point matching module is used to match feature points of two adjacent frames in the segmented video; A calculation module, used to calculate the number of mismatched point pairs of each frame image and obtain the distribution of mismatched point pairs of the segmented video, and to calculate the image quality of each frame image and obtain the image quality distribution of the segmented video; The result judgment module is used to judge the consistency between the distribution of mismatched point pairs and the image quality distribution and generate a video jitter recognition signal.

6. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the method according to any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a processor, the processor is caused to execute the method according to any one of claims 1 to 4.

Images