Adaptive serial asynchronous video comparison method
By selecting the comparison monitoring point in the video broadcast link and processing the feature value difference of video image frames, the problem of difficulty in adaptive search delay time and synchronization comparison in the prior art is solved, and the video broadcast link is realized Real-time monitoring and fault handling of the road improve the stability of the broadcast process.
Patent Information
- Application Number
- CN202411033494.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-12-06
AI Technical Summary
The prior art is difficult to adaptively find delay time and realize synchronous comparison in the video broadcast link, which makes it difficult to detect and deal with fault points in time during the video broadcasting process, which may lead to broadcast accidents.
Select two comparison monitoring points on the broadcast link. After scaling and converting the video image frame size, the feature value is obtained, the difference value is compared, and the feature value is processed in the feature value queue to synchronize the video stream and judge the image Whether the difference value of the frame is smaller than the preset value to determine the synchronous state of the video stream and trigger an alarm.
It realizes adaptive search delay time, monitors the synchronization of the video broadcast link data flow in real time, triggers alarms in a timely manner, and improves the stability of the broadcast process.
Smart Images

Figure CN119094732A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of asynchronous video comparison, and in particular to a method for adaptive serial asynchronous video comparison. Background Art
[0002] In the safe broadcast of TV programs, a set of programs must pass through several parts such as the audio and video source end, the transmission part end, the video output end, and the video viewing end. These parts are connected in series, allowing the signal to pass through each part in turn and finally be sent to the user's TV, mobile phone and other terminals. Among all the above-mentioned devices, links, and networks, all devices except the user terminal are in saturated operation 7×24 hours a week, so it is inevitable that any link will have an abnormality. Upstream errors will lead to downstream errors. Therefore, sometimes a small "point" abnormality may lead to "chain" abnormality, or even "surface" abnormality, and eventually cause the entire broadcast link to be abnormal, and in severe cases, cause broadcast accidents. Simply put, if the user sees an abnormal picture, it may be caused by a point in the above-mentioned links, but it is impossible to determine which point is the cause.
[0003] To solve this problem, the traditional detection method is to draw some detection points at the middle points of the above links and then detect these detection points. Figure 1 As shown. The "first monitoring point 3" is introduced before the "first broadcast link 1", the "second monitoring point 4" is introduced between the "first broadcast link 1" and the "second broadcast link 2", and the "third monitoring point 5" is introduced after the "second broadcast link 2". It is not difficult to imagine that if the "first monitoring point 3" is abnormal, then the upstream (i.e., input source) of the "first broadcast link 1" in the figure has an abnormality; if the "first monitoring point 3" is normal, and the "second monitoring point 4" is abnormal, then the "first broadcast link 1" in the figure has an abnormality; if the "second monitoring point 4" is normal, and the "third monitoring point 5" is abnormal, then the "second broadcast link 2" in the figure has an abnormality. In the early days, this judgment was made manually by the duty officer. Although this method is intuitive and effective, when an abnormality occurs, it is very difficult for the duty officer to quickly judge the abnormal point from the combination of various abnormalities and make normal operations. Sometimes, mistakes will be made in a hurry, leading to more errors. In the past 20 years, almost all TV stations have added monitoring systems and made judgments through software. However, this only increases the timeliness and accuracy of each monitoring point, and fails to solve the essential problem.
[0004] To compare videos, the most effective way is to compare the images of the monitoring points involved in the comparison pixel by pixel. If all are the same, it is considered normal; otherwise, it is considered abnormal. However, in reality, this method has many problems and cannot be applied to actual systems at all. The link between the two monitoring points mentioned above may be just one device, or it may be a link, or it may be a complex system. No matter which one, as long as the information is transcoded, there may be differences in the input and output signals. In layman's terms, the transcoded images may be completely consistent by the naked eye, but from the software's point of view, the resolution may have changed (such as 3840×2160 to 1920×1080), the image parameters may have changed (such as saturation increase), and the encoding method may have changed (such as RGB to YUV). Therefore, the "consistent images" mentioned above do not mean that the images must be exactly the same. As long as the images are very similar, they can be considered "consistent". In layman's terms, when comparing the pixels at the same position on the two images, "equal" cannot be used, but only "close". Therefore, the pixel-by-pixel comparison method is not feasible.
[0005] The devices, links or systems in the links may all cause time delays. For example, the above transcoding will inevitably cause delays. Sometimes, even without transcoding, delays will still occur. For example, the geosynchronous satellite is located about 36,000 kilometers above the ground. Even if the transmission speed reaches 300,000 kilometers per second, the delay error between the uplink and the receiving end is at least 240 milliseconds (3.6×2 / 30=0.24). Therefore, the "image comparison" mentioned above does not compare the images of the input and output ends at the same time point, but compares the images that appeared before the input end with the current images of the output end.
[0006] The delay time generated by different links varies greatly. Generally speaking, the transcoding server needs to delay 1000-3000 milliseconds, and as mentioned above, if the conversion is performed through a geosynchronous satellite, it will take more than 240 milliseconds. More importantly, if there is a switching device in the link, the delay time between two monitoring points on a link will change at a certain moment. Therefore, the existing algorithm cannot adaptively find the delay time. Summary of the invention
[0007] The purpose of the present invention is to provide a method for adaptive serial asynchronous video comparison to solve the above technical problems;
[0008] A method for adaptive serial asynchronous video comparison, comprising:
[0009] Step S1, selecting two comparison monitoring points on the broadcast link, and scaling and format conversion of the image frame resolution of the original video in the broadcast link to obtain the feature value corresponding to the image frame;
[0010] Step S2, comparing the feature values corresponding to adjacent image frames to obtain a difference value;
[0011] Step S3, setting a characteristic value queue at the upstream position and the downstream position of the broadcast link respectively, and performing queue configuration processing on the upstream position and the downstream position;
[0012] Step S4, if the video streams of the two compared monitoring points are not synchronized, then execute step S5; if the video streams of the two compared monitoring points are synchronized, then execute step S6;
[0013] Step S5, traversing all the difference values in the feature value queue configured at the upstream position to obtain a maximum difference value, and synchronously processing the video streams of the two compared monitoring points based on the maximum difference value and the set minimum difference threshold constant and minimum comparison threshold constant;
[0014] Step S6, comparing the upstream position and the downstream position to determine whether the difference value is less than a second preset value, if yes, determining that the two image frames are the same, if not, determining that the two image frames are different, and triggering an alarm.
[0015] Preferably, step S1 scales the image frame resolution of the original video in the broadcast link and then converts it into a grayscale image;
[0016] The resolution size after scaling is 256×256, and the characteristic value corresponding to a single image frame is an array including 65536 values of 0 to 255.
[0017] Preferably, step S2 includes subtracting the two groups of 65536 values in sequence, calculating a first absolute value of the difference after the subtraction, and counting the values of the first absolute values that are greater than a first preset value, and the statistical result represents the difference value of the two image frames;
[0018] The first preset value is 10.
[0019] Preferably, the upstream position in step S3 is the signal inflow end of the broadcast link, and the downstream position is the signal outflow end of the broadcast link;
[0020] The length of the feature value queue at the upstream position is N, and the length of the feature value queue at the downstream position is 1. Preferably, the calculation formula for the length of the feature value queue at the upstream position is:
[0021] N = t × 2 / 40;
[0022] Wherein, N represents the characteristic value queue length of the upstream position;
[0023] t represents the estimated delay time.
[0024] Preferably, the queue configuration process in step S3 includes:
[0025] When the latest image frame is received, the image frame is scaled and converted into a grayscale image to obtain the latest eigenvalue corresponding to the latest image frame, and then the latest eigenvalue and the latest difference value are inserted into the head of the eigenvalue queue, and the eigenvalue and the difference value at the tail of the eigenvalue queue are deleted.
[0026] Preferably, each item of the feature value queue in step S3 includes an array of 65536 bytes and a difference value.
[0027] Preferably, step S5 comprises,
[0028] Step S51, traversing all the difference values in the feature value queue configured at the upstream position to obtain the maximum difference value;
[0029] Step S52: if the maximum difference value is less than the minimum difference threshold constant, it is determined that the current state is not suitable for synchronization; if the maximum difference value is greater than or equal to the minimum difference threshold constant, the maximum difference value of the upstream position is subtracted from the difference value of the downstream position to obtain a second absolute value; when the second absolute value is less than the minimum comparison threshold constant, it is determined that the synchronization is completed.
[0030] Preferably, step S6 comprises,
[0031] Step S61, comparing the latest feature value of the downstream position with the feature value corresponding to the maximum difference value in the feature value queue configured at the upstream position to obtain the difference value;
[0032] Step S62: if the difference value is less than the second preset value, it is determined that the two image frames are the same; if the difference value is greater than or equal to the second preset value, it is determined that the two image frames are different, and an alarm is triggered.
[0033] Preferably, in step S5, the minimum difference threshold constant is 32767, and the minimum comparison threshold constant is 10;
[0034] In step S6, the second preset value is 65536.
[0035] The beneficial effects of the present invention are: being able to adaptively search for delay time, and to monitor the synchronization of data streams in a video broadcast link in real time, and being able to promptly trigger an alarm and notify an operator when a fault point occurs, thereby improving the stability of the broadcast process. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] Figure 1 This is a schematic diagram of the broadcast link anomaly detection in the prior art;
[0037] Figure 2 It is a step diagram of the method for adaptive serial asynchronous video comparison of the present invention;
[0038] Figure 3 is a schematic diagram of step S5 of the present invention;
[0039] Figure 4 It is a schematic diagram of step S6 of the present invention. DETAILED DESCRIPTION
[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0041] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.
[0042] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.
[0043] A method for adaptive serial asynchronous video comparison, such as Figure 2 As shown, including,
[0044] Step S1, selecting two comparison monitoring points on the broadcast link, and scaling and format conversion of the image frame resolution of the original video in the broadcast link to obtain the feature value corresponding to the image frame;
[0045] Step S2, comparing the feature values corresponding to adjacent image frames to obtain a difference value;
[0046] Step S3, setting a characteristic value queue at the upstream position and the downstream position of the broadcast link respectively, and performing queue configuration processing on the upstream position and the downstream position;
[0047] Step S4, if the video streams of the two compared monitoring points are not synchronized, then execute step S5; if the video streams of the two compared monitoring points are synchronized, then execute step S6;
[0048] Step S5, traversing all difference values in the feature value queue configured at the upstream position to obtain the maximum difference value, and synchronously processing the video streams of the two compared monitoring points based on the maximum difference value and the set minimum difference threshold constant and minimum comparison threshold constant;
[0049] Step S6, compare the upstream position and the downstream position to determine whether the difference value is less than a second preset value, if yes, determine that the two image frames are the same, if not, determine that the two image frames are different, and trigger an alarm.
[0050] Specifically, the present invention provides an adaptive serial asynchronous video comparison method, which can adaptively search for delay time and realize automatic processing of video stream synchronization and abnormal status, which not only improves the monitoring efficiency, but also reduces the workload of operators and improves the stability of the broadcast process.
[0051] In a preferred embodiment, step S1 scales the image frame resolution of the original video in the broadcast link and then converts it into a grayscale image;
[0052] The scaled resolution is 256×256, and the feature value corresponding to a single image frame is an array of 65536 values from 0 to 255.
[0053] Specifically, the scaled image frames are unified into a size of 256×256 and converted into grayscale images, so that each image frame has the same feature representation, which can simplify the subsequent feature extraction and comparison process. Because each image frame corresponds to a feature vector of a fixed size, the unified feature representation and accurate matching capability enable the system to more reliably monitor and identify synchronization problems or abnormal conditions in the broadcast link.
[0054] In a preferred embodiment, step S2 includes sequentially subtracting the two groups of 65536 values, calculating a first absolute value of the difference after the subtraction, and counting the values of the first absolute values that are greater than a first preset value, and the statistical result represents the difference value of the two image frames;
[0055] The first preset value is 10.
[0056] Specifically, in the first absolute value, those differences that are greater than a first preset value (here 10) are counted. These differences represent significant changes or differences between image frames, reflecting substantial changes in image content, and the statistical results represent the difference between two image frames, which can be used to determine changes in the video stream.
[0057] Specifically, the setting of the first preset value affects the sensitivity to the difference. A lower preset value may make it easier to detect subtle changes, while a higher preset value may filter out some unimportant differences, which helps the system to perform accurate anomaly detection in different application scenarios.
[0058] In a preferred embodiment, the upstream position in step S3 is the signal inflow end of the broadcast link, and the downstream position is the signal outflow end of the broadcast link;
[0059] The length of the feature value queue at the upstream position is N, and the length of the feature value queue at the downstream position is 1.
[0060] Specifically, the length of the feature value queue at the upstream position is N, and the length of the feature value queue at the downstream position is fixed to 1, which can help balance and adjust the signal flow in the broadcast link. If there are multiple signal sources or complex signal processing processes upstream, by setting the feature value queue with a length of N, the signals of multiple input sources can be buffered and processed, thereby ensuring the stability and throughput of the upstream link.
[0061] In a preferred embodiment, the calculation formula for the feature value queue length at the upstream position is:
[0062] N = t × 2 / 40;
[0063] Where N represents the length of the feature value queue at the upstream location;
[0064] t represents the estimated delay time.
[0065] In a preferred embodiment, the queue configuration process in step S3 includes:
[0066] When the latest image frame is received, the image frame is scaled and converted into a grayscale image to obtain the latest eigenvalue corresponding to the latest image frame, and then the latest eigenvalue and the latest difference value are inserted into the head of the eigenvalue queue, and the eigenvalue and difference value at the tail of the eigenvalue queue are deleted;
[0067] Each item of the feature value queue in step S3 includes an array of 65536 bytes and a difference value.
[0068] Specifically, in the two monitoring points participating in the comparison (i.e., the comparison monitoring points), a first-in-first-out (FIFO) feature value queue is set in the monitoring point at the upstream position. After extracting each latest frame of video, the feature value T0 of the frame is first calculated, and the feature value T0 of the latest frame is compared with the feature value T1 of the latest head in the feature value queue (the feature value of the previous frame of the latest frame) to obtain the difference value D (D=T0-T1) between the two. Then, the feature value T0 of the latest frame and the difference value D between the feature value of the frame and the feature value of the previous frame are added to the head of the queue together, and the feature value of the oldest frame at the tail of the queue is deleted.
[0069] More specifically, at the downstream position of the two monitoring points involved in the comparison, the configuration is the same as that of the upstream, except that there is only one set of data to be saved in the downstream feature value queue. That is, after each latest frame of video is extracted downstream, the feature value T0 of the frame is first calculated using the same algorithm, and the feature value T0 of the latest frame is compared with the unique feature value T1 in the feature value queue to obtain the difference between the two (D0=T0-T1). If the current system is not synchronized, then a synchronization operation is performed, and the time difference F between the upstream and downstream is obtained in frames. If the same frequency is not required, the feature value T0 of the downstream frame is compared with the feature value TD of the upstream frame that is offset forward by D frames. If the two are the same, the video is considered normal; if they are not the same, the video is considered abnormal. Subsequently, the feature value T0 of the current frame is overwritten with the only feature value saved in the feature value queue.
[0070] In a preferred embodiment, referring to Figure 3 , step S5 comprises,
[0071] Step S51, traversing all difference values in the feature value queue configured at the upstream position to obtain the maximum difference value;
[0072] Step S52: If the maximum difference value is less than the minimum difference threshold constant, it is determined that the current state is not suitable for synchronization; if the maximum difference value is greater than or equal to the minimum difference threshold constant, the maximum difference value of the upstream position is subtracted from the difference value of the downstream position to obtain a second absolute value; when the second absolute value is less than the minimum comparison threshold constant, it is determined that the synchronization is completed.
[0073] Specifically, the synchronization operation first presets a minimum difference threshold constant dThreshold and a minimum comparison threshold constant cThreshold. Then traverse the upstream feature value queue, find the difference value D therein, and select the maximum maximum difference dMax. Compare this value with dThreshold. If it is less than, ignore the synchronization operation, and dMax is invalid at this time. If it is greater than or equal to, continue the synchronization operation, and compare this value with the downstream difference value D. If the difference between the two is large (|upstream dMax-downstream D|>=cThreshold), ignore it. Otherwise, the position of dMax is the time difference F between the upstream and downstream, in frames.
[0074] Specifically, the practical significance of the synchronization algorithm is that the upstream video is constantly changing, so the eigenvalues of the eigenvalue queue are also different, and thus the difference values D of two adjacent frames are also different. The largest one is selected, that is, the position and difference of the two frames with the largest difference between adjacent frames in the previous video are found. If the difference is less than the minimum difference threshold dThreshold (such as during the news broadcast, only the host's mouth is moving, and the other pictures are static. Although the picture is constantly changing, the overall change of the picture is not large, so do not perform synchronization at this time), until a larger difference appears (for example, the previous frame is a serious studio news broadcast scene, and the next frame becomes a dazzling advertisement), then continue synchronization. Sometimes, in the same video, there will be multiple large-scale image switching (i.e., there are multiple points greater than the minimum difference threshold constant dThreshold), and the downstream will also appear in turn. To find the largest point, the downstream difference value D must not only be greater than the minimum difference threshold constant dThreshold, but also the same as the upstream maximum value dMax to be considered a true synchronization value. Otherwise, it can only be that the downstream has also experienced a large-scale switching, but this large-scale switching is only one of the most recent large-scale switching, not the large-scale switching corresponding to the upstream dMax. Similarly, as mentioned above, even if the upstream and downstream algorithms are exactly the same, because the upstream and downstream videos will change, when the upstream and downstream perform the same operation, they cannot be judged as equal in the mathematical sense, but should be understood as "the absolute value of the difference after subtracting the two is less than a certain preset very small threshold value".
[0075] In a preferred embodiment, referring to Figure 4 , step S6 comprises,
[0076] Step S61, comparing the latest feature value at the downstream position with the feature value corresponding to the maximum difference value in the feature value queue configured at the upstream position to obtain a difference value;
[0077] Step S62: if the difference value is less than the second preset value, the two image frames are determined to be the same; if the difference value is greater than or equal to the second preset value, the two image frames are determined to be different, and an alarm is triggered.
[0078] Specifically, if the current synchronization has been completed, the upstream and downstream can be directly compared. The algorithm is: use the feature value of the latest frame of the downstream (i.e., 65535 gray values of 0 to 255) to compare with the feature value corresponding to the frame in the item pointed to by the maximum difference value Dmax in the upstream feature value queue. The comparison algorithm is also to subtract the two groups of 65536 values in sequence and calculate the absolute value of the difference. Then count the values greater than the preset value T1 (default T1 = 10) among these 65535 absolute values, and the statistical result is D. If the value is less than 65536, the two image frames are considered to be the same. Otherwise, they are different. For the latter, an alarm can be issued so that the on-duty personnel can check whether there is a device failure.
[0079] In a preferred embodiment, in step S5, the minimum difference threshold constant is 32767 and the minimum comparison threshold constant is 10;
[0080] In step S6, the second preset value is 65536.
[0081] Specifically, in the feature value queue configured at the upstream position, all difference values are traversed and the maximum difference value is extracted. This ensures that the most significant difference between the two comparison monitoring points can be obtained, and even small changes can be accurately captured. The minimum difference threshold constant and the minimum comparison threshold constant are used to filter and process the differences.
[0082] In further detail, the minimum difference threshold constant is used to determine whether the difference between the two video streams exceeds a higher threshold. If the maximum difference value is below this threshold, it may mean that the two video streams are very similar or almost the same. The minimum comparison threshold constant is used to compare the maximum difference value with the set second preset value, so that even a small difference can be detected, thereby avoiding underreporting.
[0083] More specifically, the video frames at the upstream position and the downstream position are compared to determine whether the difference value between them is less than a second preset value. The setting of this threshold can help determine whether two image frames are visually similar. If the difference value is less than the second preset value, the system determines that the two image frames are the same and no further processing or triggering is required; if the difference value is greater than the second preset value, the system determines that the two image frames are different and triggers an alarm. False alarms can be reduced and the reliability and practicality of the system can be improved.
[0084] The specific implementation steps of the method for adaptive serial asynchronous video comparison provided by the present invention are as follows:
[0085] Step 1: For each image frame, convert it into a 256×256 grayscale array as the feature value of the image.
[0086] Step 2: queue cache the upstream, and make a difference value D for two adjacent frames in the upstream. The algorithm is to subtract the two sets of 65536 values in sequence and calculate the absolute value of the difference. Then count the values of these 65535 absolute values that are greater than the preset value T1 (default T1 = 10). For the downstream, only the difference value D between the current frame and the previous frame is calculated, and the algorithm is the same as the upstream.
[0087] Step 3: If there is no synchronization, all the difference values D in the upstream feature value queue are traversed to find the maximum value dMax. If this value is less than the minimum difference threshold dThreshold, it is invalid. Otherwise, this value is compared with the downstream difference value D. If the value F is equal to the minimum comparison threshold cThreshold constant, then F is the number of frames between the asynchronous upstream and downstream.
[0088] Step 4: In the upstream cache, the feature value of the frame pointed to by F is compared with the feature value of the current frame downstream. If the two are the same, the video is considered normal; if they are different, the video is considered abnormal. The same algorithm here is: use the feature value of the latest frame downstream (i.e. 65535 grayscale values 0 to 255) and compare it with the feature value corresponding to the frame with the maximum difference value F in the upstream feature value queue (i.e. the item pointed to by dMax). The comparison algorithm is also to subtract the two groups of 65536 values in sequence and calculate the absolute value of the difference. Then count the values greater than the preset value T1 (default T1 = 10) among these 65535 absolute values, and the statistical result is D. If the value is less than 65536, the two image frames are considered the same. Otherwise, they are different.
[0089] In summary, the present application provides a method for adaptive serial asynchronous video comparison, in which the characteristic values of historical images are cached in a queue manner at the upstream node of the device or link. If the number of frames of asynchronous phase difference needs to be calculated, the maximum jump position and jump value of the buffer image also need to be calculated. Only the characteristic value of the previous frame of historical image is cached at the downstream node of the device or link. When the number of frames of asynchronous phase difference needs to be calculated, it can be obtained through the maximum jump position and jump value of the upstream and downstream. When the number of frames of asynchronous phase difference does not need to be calculated, it is only necessary to directly compare the characteristic value of the current image of the current downstream with the image characteristic value of the frame pointed to by the maximum jump position of the upstream to obtain the comparison result of the upstream and downstream asynchronous videos, and then determine whether to alarm. After field testing, this method can well compare serial asynchronous videos in devices, links, and even a network, and has an adaptive effect.
[0090] The above description is only a preferred embodiment of the present invention, and does not limit the implementation mode and protection scope of the present invention. For those skilled in the art, it should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the description and illustrations of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for adaptive serial asynchronous video comparison, characterized in that: include, Step S1, selecting two comparison monitoring points on the broadcast link, and scaling and format conversion of the image frame resolution of the original video in the broadcast link to obtain the feature value corresponding to the image frame; Step S2, comparing the feature values corresponding to adjacent image frames to obtain a difference value; Step S3, setting a characteristic value queue at the upstream position and the downstream position of the broadcast link respectively, and performing queue configuration processing on the upstream position and the downstream position; Step S4, if the video streams of the two compared monitoring points are not synchronized, then execute step S5; if the video streams of the two compared monitoring points are synchronized, then execute step S6; Step S5, traversing all the difference values in the feature value queue configured at the upstream position to obtain a maximum difference value, and synchronously processing the video streams of the two compared monitoring points based on the maximum difference value and the set minimum difference threshold constant and minimum comparison threshold constant; Step S6, comparing the upstream position and the downstream position to determine whether the difference value is less than a second preset value, if yes, determining that the two image frames are the same, if not, determining that the two image frames are different, and triggering an alarm.
2. The method for adaptive serial asynchronous video comparison according to claim 1, characterized in that: Step S1: scaling the image frame resolution of the original video in the broadcast link and then converting it into a grayscale image; The resolution size after scaling is 256×256, and the characteristic value corresponding to a single image frame is an array including 65536 values of 0 to 255.
3. The method for adaptive serial asynchronous video comparison according to claim 1, characterized in that: Step S2 includes subtracting the two groups of 65536 values in sequence, calculating a first absolute value of the difference after the subtraction, and counting the values of the first absolute values that are greater than a first preset value, and the statistical result represents the difference value of the two image frames; The first preset value is 10.
4. The method for adaptive serial asynchronous video comparison according to claim 1, characterized in that: In step S3, the upstream position is the signal inflow end of the broadcast link, and the downstream position is the signal outflow end of the broadcast link; The length of the feature value queue at the upstream position is N, and the length of the feature value queue at the downstream position is 1.
5. The method for adaptive serial asynchronous video comparison according to claim 4, characterized in that: The calculation formula of the characteristic value queue length of the upstream position is: N = t × 2 / 40; Wherein, N represents the characteristic value queue length of the upstream position; t represents the estimated delay time.
6. The method for adaptive serial asynchronous video comparison according to claim 1, characterized in that: The queue configuration process in step S3 includes: When the latest image frame is received, the image frame is scaled and converted into a grayscale image to obtain the latest eigenvalue corresponding to the latest image frame, and then the latest eigenvalue and the latest difference value are inserted into the head of the eigenvalue queue, and the eigenvalue and the difference value at the tail of the eigenvalue queue are deleted.
7. The method for adaptive serial asynchronous video comparison according to claim 1, characterized in that: Each item of the feature value queue in step S3 includes an array of 65536 bytes and a difference value.
8. The method for adaptive serial asynchronous video comparison according to claim 1, characterized in that: Step S5 comprises, Step S51, traversing all the difference values in the feature value queue configured at the upstream position to obtain the maximum difference value; Step S52: if the maximum difference value is less than the minimum difference threshold constant, it is determined that the current state is not suitable for synchronization; if the maximum difference value is greater than or equal to the minimum difference threshold constant, the maximum difference value of the upstream position is subtracted from the difference value of the downstream position to obtain a second absolute value; when the second absolute value is less than the minimum comparison threshold constant, it is determined that the synchronization is completed.
9. The method for adaptive serial asynchronous video comparison according to claim 1, characterized in that: Step S6 comprises, Step S61, comparing the latest feature value of the downstream position with the feature value corresponding to the maximum difference value in the feature value queue configured at the upstream position to obtain the difference value; Step S62: if the difference value is less than the second preset value, it is determined that the two image frames are the same; if the difference value is greater than or equal to the second preset value, it is determined that the two image frames are different, and an alarm is triggered.
10. The method for adaptive serial asynchronous video comparison according to claim 1, characterized in that: In step S5, the minimum difference threshold constant is 32767, and the minimum comparison threshold constant is 10; In step S6, the second preset value is 65536.
