Low-latency video security exchange method for video surveillance center

Abstract

The application relates to the technical field of video security exchange, in particular to a low-delay video security exchange method suitable for a video monitoring center, which comprises the following steps: collecting each frame of monitoring video, and uniformly dividing each frame of monitoring video; determining the spatial domain complexity of each image block of each frame of monitoring video; obtaining the frequency domain complexity of each image block of each frame of monitoring video; combining the spatial domain complexity and the frequency domain complexity to determine the comprehensive content complexity of each image block of each frame of monitoring video, so as to modify the quantization parameter when each image block of each frame of monitoring video is video encoded; and performing security exchange on each frame of encoded monitoring video. The application realizes accurate judgment on the content complexity of monitoring video and self-adaptive encoding, and improves the low-delay effect of video security exchange of the video monitoring center.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of video security exchange, in particular to a low-latency video security exchange method suitable for a video monitoring center. BACKGROUND

[0002] In traffic management, criminal investigation and other scenarios, the monitoring information obtained by the video monitoring center plays an important role in maintaining public safety. For the video monitoring center, the real-time and accuracy of video information are crucial, and how to realize low-latency transmission while ensuring video quality has become an important direction of current video monitoring technology research.

[0003] The encoding method based on content complexity dynamically adjusts the encoding parameters by analyzing the content complexity of the image frames, ensures the quality of high complexity pictures, and processes low complexity pictures with higher compression encoding parameters to improve transmission efficiency. However, in actual monitoring scenarios, the influence of the external environment will interfere with the accuracy of the picture complexity judgment. When the monitoring device is affected by wind or vibration and appears to be shaking, the adjacent frame monitoring video will show a large change in pixel value and motion vector, thus being misjudged as a high complexity scene, resulting in that the monitoring picture with low importance occupies too high code rate, and finally affecting the real-time performance of video exchange. SUMMARY

[0004] In order to solve the above technical problems, the present application provides a low-latency video security exchange method suitable for a video monitoring center to solve the existing problems.

[0005] The low-latency video security exchange method suitable for a video monitoring center of the present application adopts the following technical scheme:

[0006] One embodiment of the present application provides a low-latency video security exchange method suitable for a video monitoring center, which comprises the following steps:

[0007] Collecting each frame of monitoring video, uniformly dividing each frame of monitoring video;

[0008] Analyzing the gray difference of the pixel points between the corresponding position image blocks in each frame of monitoring video and its adjacent frame of monitoring video, and the motion change information of each frame of monitoring video relative to its adjacent frame of monitoring video, to determine the spatial complexity of each image block of each frame of monitoring video;

[0009] Performing frequency domain transformation on each frame of monitoring video, determining the frequency domain complexity of each image block of each frame of monitoring video through the amplitude difference of the frequency domain points between the corresponding position image blocks in each frame of monitoring video and its adjacent frame of monitoring video, and the linear change degree of the phase difference between the corresponding position image blocks;

[0010] Determine a comprehensive content complexity of each image block of each frame of the monitoring video by combining the spatial complexity and the frequency domain complexity, so as to modify a quantization parameter when encoding each image block of each frame of the monitoring video;

[0011] Perform a secure exchange of the encoded frames of the monitoring video.

[0012] In one embodiment, the determination of the spatial complexity of each image block of each frame of the monitoring video comprises:

[0013] Calculate a mean value of a gray scale difference of all pixel points between each image block of each frame of the monitoring video and a corresponding image block of a previous frame of the monitoring video; and determine an optical flow vector of each feature point in each image block of each frame of the monitoring video by using an optical flow method.

[0014] Determine a displacement variation degree of each image block of each frame of the monitoring video based on a numerical distribution of a modulus of the optical flow vector.

[0015] The spatial complexity is positively correlated with the mean value and the displacement variation degree.

[0016] In one embodiment, the displacement variation degree is a fusion result of a mean value and a dispersion degree of the modulus of the optical flow vector of each feature point in each image block of each frame of the monitoring video.

[0017] In one embodiment, the spatial complexity is a product of the mean value and the displacement variation degree.

[0018] In one embodiment, the determination of the frequency domain complexity of each image block of each frame of the monitoring video comprises:

[0019] Calculate a fusion result of an amplitude difference of all frequency domain points between each image block of each frame of the monitoring video and a corresponding image block of a previous frame of the monitoring video, denoted as a first fusion value; and calculate a fusion result of a square of an amplitude of all frequency domain points of the corresponding image block of the previous frame of the monitoring video, denoted as a second fusion value.

[0020] Determine an amplitude difference degree of each image block of each frame of the monitoring video based on the first fusion value and the second fusion value.

[0021] Calculate a phase difference of all frequency domain points between each image block of each frame of the monitoring video and a corresponding image block of a previous frame of the monitoring video, perform a linear fitting on the phase difference to obtain a residual term of the linear fitting, calculate a fusion result of a square of all the residual terms between each image block of each frame of the monitoring video and the corresponding image block of the previous frame of the monitoring video, denoted as a third fusion value.

[0022] The frequency domain complexity is positively correlated with the amplitude difference degree and the third fusion value.

[0023] In one embodiment, the further determining the amplitude difference degree comprises:

[0024] calculating a sum of the second fusion value and a preset value greater than 0, and the amplitude difference degree is a ratio of the first fusion value and the sum.

[0025] In one embodiment, the frequency domain complexity is a product of the amplitude difference degree and the third fusion value.

[0026] In one embodiment, the determining the comprehensive content complexity of each image block of each frame of the monitoring video comprises:

[0027] If the normalized value of the spatial domain complexity of any image block of each frame of the monitoring video is less than or equal to a preset threshold, the comprehensive content complexity of the any image block is the normalized value of the spatial domain complexity of the image block; otherwise, the comprehensive content complexity of the any image block is a product of the normalized value of the spatial domain complexity and the normalized value of the frequency domain complexity of the image block.

[0028] In one embodiment, the correcting the quantization parameter in the video encoding of each image block of each frame of the monitoring video comprises:

[0029] presetting a maximum value and a minimum value of the quantization parameter, calculating a difference between the maximum value and the minimum value, and determining the corrected quantization parameter in the video encoding of each image block of each frame of the monitoring video by combining the comprehensive content complexity of each image block of each frame of the monitoring video.

[0030] In one embodiment, the corrected quantization parameter is a difference between the maximum value and a multiplication result of the comprehensive content complexity of each image block of each frame of the monitoring video and the difference.

[0031] The present application has at least the following beneficial effects:

[0032] The present application can accurately identify the detail degree and change of each image block in each frame of monitoring video by collecting each frame of monitoring video, uniformly dividing each frame of monitoring video, determining the spatial complexity of each image block of each frame of monitoring video, improving the accuracy of image block processing, so that the details of each image block have higher control in the video coding process, and the controllability of video quality is enhanced; further, the frequency domain complexity of each image block of each frame of monitoring video is determined, which is helpful to identify the area with large frequency change in the monitoring video, so as to effectively capture the dynamic characteristics of the monitoring video, effectively avoid the loss of important frequency information in the compression process, reduce the frequency domain distortion and improve the video quality; through the comprehensive analysis of the spatial complexity and the frequency domain complexity of each frame of monitoring video, the quantization parameter of each image block can be more accurately adjusted, so that the coding process is more intelligent, which helps to ensure that each image block adopts the most suitable coding strategy according to the complexity of its content, so as to optimize the coding quality and reduce the redundant information, improve the accuracy and flexibility of video coding, and reduce unnecessary data amount during coding, so that the video data transmission speed is accelerated, thereby reducing the delay, enhancing the low delay of video monitoring exchange, the present application can flexibly adapt to frequently changing scenes, automatically adjust the coding parameters, solve the coding efficiency problem when the scene changes frequently, and help to keep the consistency of coding effect for the complex scene switching often occurring in the monitoring environment, avoid the coding distortion caused by scene change, and finally avoid the misjudgment of the shaking picture of the monitoring device as a picture with high content complexity, realize the accurate judgment of the content complexity of the monitoring video and the adaptive coding, and improve the low delay effect of the video security exchange of the video monitoring center. BRIEF DESCRIPTION OF DRAWINGS

[0033] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings needed in the following embodiment or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0034] Figure 1 The step flow chart of the low delay video security exchange method for the video monitoring center provided by the present application;

[0035] Figure 2 The adaptive QP value determination flow chart for image block coding. DETAILED DESCRIPTION

[0036] For further elaboration of the technical means and effects taken by the present application to achieve the predetermined object, the specific implementation, structure, features and effects of the low-latency video security exchange method for video monitoring center according to the present application are described in detail as follows in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0037] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0038] The specific scheme of the low-latency video security exchange method for video monitoring center provided by the present application is described in detail below in combination with the drawings.

[0039] The low-latency video security exchange method for video monitoring center provided by one embodiment of the present application, in particular, provides the following low-latency video security exchange method for video monitoring center, please refer to Figure 1 , the method comprises the following steps:

[0040] Step S001, collecting each frame of monitoring video, and uniformly dividing each frame of monitoring video.

[0041] The video picture of the monitoring scene is collected in real time by the monitoring camera to obtain scene video data. The collected video data is first uniformly processed in format, and the original video stream is uniformly converted into RGB data format to ensure the compatibility and efficiency of subsequent processing.

[0042] The continuous T frames of monitoring video are stored in the cache queue by using the frame buffering mechanism, which is used to analyze the difference between adjacent frames and the content complexity. The size of T can be set by the implementer according to the implementation scene, and is not specially limited, and in this embodiment, T=5. The frame buffering mechanism is a known technology, and its specific process is not described in detail.

[0043] The first frame of monitoring video in the cache queue is represented as , and first, each frame of monitoring video is divided into blocks, and the size of each image block can be set by the implementer according to the real-time scene, and is not specially limited. If more detailed information needs to be processed, it can be adjusted to be smaller, but it should meet the macro block size standard in H.265 coding, and in this embodiment ​The size of each image block is 32*32, and each image block is obtained by dividing each frame of monitoring video. For an image block with a size less than 32*32, a mirror filling method is used for filling. The mirror filling method is a known technology, and thus the process is not described herein.

[0044] In step S002, the gray level difference between the pixel points of each frame of monitoring video and the corresponding image block of the adjacent frame of monitoring video is analyzed, and the motion change information of each frame of monitoring video relative to the adjacent frame of monitoring video is determined, so as to determine the spatial complexity of each image block of each frame of monitoring video.

[0045] For a monitoring scene affected by external environmental factors, the monitoring video screen may be shaken. For example, the monitoring device is affected by wind or vibration. When the content complexity of an image screen is analyzed by using a traditional video encoding method based on content complexity, the complexity of the screen is determined by relying on the change of adjacent image frames on the pixel, which may interfere with the accuracy. For example, when the monitoring screen is shaken, the pixel points of continuous frames of monitoring video are greatly different, and the motion vector of the pixel points is also large, so that the monitoring screen without substantial content change is misjudged as a high complexity scene screen, the encoder reduces the encoding compression ratio of the video screen, and a high resource is occupied in transmission, which affects the low delay performance of the entire video transmission.

[0046] Based on the above analysis, the present application considers the features of continuous different frames of monitoring video in the spatial domain and the frequency domain, obtains the comprehensive content complexity of the monitoring video by the two features, encodes the monitoring video, realizes accurate judgment of the complexity of the monitoring video and adaptive encoding, and improves the low delay effect of the video security exchange of the subsequent video monitoring center.

[0047] Specifically, the spatial domain features of each frame of monitoring video are analyzed first. For each frame of monitoring video divided into blocks, the average pixel difference between the corresponding image blocks of the t-th frame of monitoring video and the previous frame of monitoring video is calculated. The specific expression is as follows: wherein, is the average pixel difference of the n-th image block of the t-th frame of monitoring video, M is the number of pixel points in each image block, and M=32*32 in the present embodiment, is the gray value of the i-th pixel point in the n-th image block of the t-th frame of monitoring video, is the gray value of the i-th pixel point in the n-th image block of the t-1-th frame of monitoring video.

[0048] It should be noted that, ​The average pixel difference is used to represent the gray difference between the ith pixel point in the nth image block of the tth frame of monitoring video and the ith pixel point in the nth image block of the (t-1)th frame of monitoring video, wherein the difference represents the difference between two variables, and can be calculated by using the absolute value of the difference, the square of the difference, the ratio, etc., which is not limited in the embodiment. The average pixel difference is used to represent the gray difference between the ith pixel point in the nth image block of the tth frame of monitoring video and the ith pixel point in the nth image block of the (t-1)th frame of monitoring video, wherein the difference represents the difference between two variables, and can be calculated by using the absolute value of the difference, the square of the difference, the ratio, etc., which is not limited in the embodiment. The greater the average pixel difference is, the greater the change of the content in the monitoring video is, and the object in the monitoring scene can change or the scene brightness can change.

[0049] Further, based on the tth frame of monitoring video and the (t-1)th frame of monitoring video, the optical flow method is used to determine the optical flow vector of each feature point in each image block of the tth frame of monitoring video. Then, the average value of the optical flow vector module length of all feature points in the nth image block is calculated. The greater the average value is, the greater the motion information of the object in the nth image block is. The dispersion degree is calculated by using the variance, the standard deviation, the coefficient of variation, etc., and the variance is used as the calculation method of the dispersion degree in the embodiment.

[0050] The fusion result of the average value and the dispersion degree is used as the displacement change degree of the nth image block of the tth frame of monitoring video. The greater the displacement change degree is, the greater the displacement change of the object in the nth image block is. The greater the displacement change degree is, the greater the displacement change of the object in the nth image block is.

[0051] It should be noted that the fusion represents the combination of multiple variables, which can be calculated by using the addition, the multiplication, the addition-multiplication hybrid, the average calculation, etc., which is not limited in the embodiment.

[0052] The expression of the displacement change degree of the nth image block of the tth frame of monitoring video in the embodiment is as follows: The greater the displacement change degree is, the greater the displacement change of the object in the nth image block is. The optical flow method for calculating the optical flow vector of the monitoring video is a prior art, and the specific process is not described herein. The average pixel difference is used to represent the gray difference between the ith pixel point in the nth image block of the tth frame of monitoring video and the ith pixel point in the nth image block of the (t-1)th frame of monitoring video, wherein the difference represents the difference between two variables, and can be calculated by using the absolute value of the difference, the square of the difference, the ratio, etc., which is not limited in the embodiment. The greater the average pixel difference is, the greater the change of the content in the monitoring video is, and the object in the monitoring scene can change or the scene brightness can change.

[0053] Based on the average pixel difference and the displacement change degree, the spatial complexity of the nth image block of the tth frame of monitoring video is determined, wherein the spatial complexity is positively correlated with the average pixel difference and the displacement change degree.

[0054] In the embodiment, the expression of the spatial complexity of the nth image block of the tth frame of monitoring video is as follows: The greater the spatial complexity is, the greater the displacement change of the object in the nth image block is. The spatial complexity comprehensively considers the change degree of pixel value and the motion information of pixel point. If the spatial complexity is larger, it indicates that the image block has a larger change on the picture and has a higher content complexity. Conversely, it indicates that the change in the image block is smaller and is mostly static background, and the content complexity is lower.

[0055] In step S003, the frequency domain complexity of each image block of each frame of monitoring video is determined by the amplitude difference and the linear change degree of phase difference of the frequency domain points between the corresponding position image blocks in each frame of monitoring video and its adjacent frame of monitoring video.

[0056] When the monitoring device is affected by wind or vibration, the monitoring picture will be shaken, and the adjacent frame of monitoring picture will also show a larger difference in pixel point characteristics and a stronger motion feature, and a higher spatial complexity. As a result, the original static background picture is misjudged as a picture with high content complexity. Based on this, the embodiment further considers the features of the continuous frames of monitoring video in the frequency domain to construct the frequency domain complexity and comprehensively judge the content complexity of the monitoring video.

[0057] Specifically, Fourier transform is performed on each image block of each frame of monitoring video to transform the image block from the spatial domain to the frequency domain and obtain the amplitude and phase information. The Fourier transform is a known technology, and its specific process will not be described again.

[0058] The amplitude information of the mth frequency domain point on the nth image block of the tth frame of monitoring video is denoted as A (m, n, t), and the phase information is denoted as φ (m, n, t). The amplitude information of the mth frequency domain point on the nth image block of the t-1th frame of monitoring video is denoted as A (m, n, t-1), and the phase information is denoted as φ (m, n, t-1). The amplitude information of the mth frequency domain point on the nth image block of the t-1th frame of monitoring video is denoted as A (m, n, t-1), and the phase information is denoted as φ (m, n, t-1). The amplitude information of the mth frequency domain point on the nth image block of the t-1th frame of monitoring video is denoted as A (m, n, t-1), and the phase information is denoted as φ (m, n, t-1). The amplitude information of the mth frequency domain point on the nth image block of the t-1th frame of monitoring video is denoted as A (m, n, t-1), and the phase information is denoted as φ (m, n, t-1). The amplitude information of the mth frequency domain point on the nth image block of the t-1th frame of monitoring video is denoted as A (m, n, t-1), and the phase information is denoted as φ (m, n, t-1). The amplitude information of the mth frequency domain point on the nth image block of the t-1th frame of monitoring video is denoted as A (m, n, t-1), and the phase information is denoted as φ (m, n, t-1). The amplitude information of the mth frequency domain point on the nth image block of the t-1th frame of monitoring video is denoted as A (m, n, t-1), and the phase information is denoted as φ (m, n, t-1). The amplitude information of the mth frequency domain point on the nth image block of the t-1th frame of monitoring video is denoted as A (m, n, t-1), and the phase information is denoted as φ (m, n, t-1).

[0059] Taking a road monitoring picture as an example, if the monitoring picture changes due to the shaking of the monitoring device, and there is no vehicle or pedestrian passing through in the actual scene, the change of the monitoring video is mainly global translation. According to the Fourier translation theorem, the translation operation of an image will only cause a linear change in the phase spectrum in the frequency domain, and has almost no effect on the amplitude spectrum. Therefore, when the change of the monitoring picture is only caused by the shaking of the device, the difference in the frequency spectrum amplitude of the two frames of monitoring video is very small. Even if the two frames of monitoring video show a dramatic change in pixel value in the spatial domain, in the frequency domain, the energy distribution of the amplitude spectrum remains stable, there is no obvious energy migration, and the overall energy distribution structure remains unchanged.

[0060] Conversely, when vehicles or pedestrians pass through the monitored area, the movement of the objects introduces new edge texture features to the image blocks. These changes, after Fourier transform, lead to alterations in the distribution of local energy in the amplitude spectrum. In other words, as an object moves, the changes in its edge texture features are manifested in the amplitude spectrum undergoing local energy shifts between consecutive frames.

[0061] Based on the above analysis, considering the amplitude spectrum characteristics of image blocks in the frequency domain of two consecutive frames of surveillance video, the amplitude difference of each image block in each frame of surveillance video is constructed, and the specific expression is as follows:

[0062] In the formula, Let be the amplitude difference of the nth image block in the t-th frame of the surveillance video. To ensure that the value is greater than 0 and to avoid a denominator of 0, this embodiment... The implementer can set it according to the actual situation; this embodiment does not impose any restrictions on this. This is recorded as the first fusion value. This is denoted as the second fusion value.

[0063] It should be understood that amplitude difference is used to characterize the difference in energy distribution in the frequency domain amplitude spectrum between two consecutive frames of surveillance video. When the value is small, it indicates that the nth image block in the two frames of surveillance video has almost the same frequency domain energy distribution, the scene structure is basically stable, and there are no new edge texture features; when A larger value indicates a significant change in the energy distribution between image blocks in two frames of surveillance video, usually caused by the movement of objects, resulting in a significant increase in the difference in amplitude spectrum.

[0064] When a surveillance video experiences a global shift solely due to camera shake, according to Fourier's translation theorem, this global displacement will cause a change in the phase spectrum. However, this change is regular, exhibiting a global linear change along the direction of the shake. In contrast, when dynamic targets such as vehicles pass through the monitored scene, changes in the edge texture features of the video image will also cause a change in the phase spectrum, generally manifesting as local phase abrupt changes, and these changes tend to concentrate along the direction of motion.

[0065] Based on the above analysis, under the condition of monitoring equipment jitter, the phase spectrum change is a global linear shift, while under the condition of real object motion, the phase spectrum exhibits local nonlinear abrupt changes. This difference can help determine the actual complexity of the monitoring video content. Therefore, this embodiment calculates the phase difference of each frequency domain point in each image block of each frame of the monitoring video, with the specific expression as follows: In the formula, For the nth image block of the t-th frame of the surveillance video, the first... The phase difference at each frequency domain point is linearly fitted to the phase difference of all frequency domain points in the nth image block of the t-th frame of the surveillance video using the least squares plane fitting method, resulting in the linear fitting term. Based on linear fitting term Calculate the residual term The expression is: , The residual term represents the degree to which the phase difference of all frequency domain points in the nth image block of the t-th frame of the surveillance video follows a linear change.

[0066] In summary, considering the phase spectrum characteristics of image blocks in the frequency domain of two consecutive frames of surveillance video, a phase nonlinearity residual index for the nth image block of the t-th frame of surveillance video is constructed. , The nonlinearity characterizing the phase features in image blocks between two consecutive frames of surveillance video, if it represents the jitter of the surveillance equipment, then Smaller. Phase nonlinear residual index The expression is: .Will This is denoted as the third fusion value.

[0067] It should be understood that when A smaller value indicates that the phase change in this image patch remains essentially linear, which could correspond to a long-term static background in the scene or a linear phase change caused by the jitter of the monitoring equipment; conversely, when... A larger value indicates that the presence of objects in the scene has introduced new texture features, causing abrupt changes in phase in different directions.

[0068] By combining the amplitude difference degree and the phase nonlinearity residual index, the frequency domain complexity of each image block in each frame of the monitoring video is determined, wherein the frequency domain complexity is positively correlated with both the amplitude difference degree and the phase nonlinearity residual index.

[0069] In this embodiment, the t-th frame of the monitored video is the t-th frame. Frequency domain complexity of image patches The expression is: .

[0070] when A higher value indicates that the image patch has undergone significant changes in both amplitude and phase spectra due to changes in local edge and texture features caused by moving objects in the image. Conversely, if the monitored image remains unchanged or is affected by vibrations in the monitoring equipment, then... Maintaining a low level of complexity, determining the frequency domain complexity can help assess the true complexity of the surveillance footage by analyzing the spatial domain characteristics of the video.

[0071] Step S004: Combine the spatial complexity and the frequency complexity to determine the comprehensive content complexity of each image block in each frame of the surveillance video, so as to correct the quantization parameters when encoding each image block in each frame of the surveillance video.

[0072] Taking into account both the spatial and frequency domain complexity of each image patch in each frame of the surveillance video, the comprehensive content complexity of each image patch is constructed, and the specific expression is as follows: Represented as:

[0073] In the formula, Let be the overall content complexity of the nth image block in the t-th frame of the surveillance video. The Sigmoid normalization function maps spatial and frequency domain complexity to the range [0,1], facilitating its use as a weight for adaptive encoding of video content complexity. For a preset threshold, and In this embodiment The smaller the threshold, the more timely the frequency domain complexity can be introduced to assist in judging the true complexity of the surveillance video; the larger the threshold, the more effectively the low-latency performance of video switching can be ensured.

[0074] The calculation of the overall content complexity can be performed when the spatial complexity exceeds a threshold. In this case, frequency domain complexity is introduced to make a second determination of the true complexity of the image block. This ensures that the complexity of the image block will not be misjudged due to the influence of image jitter, and also avoids the waste of resources caused by performing frequency domain calculations every time, which would affect the low-latency performance of video switching.

[0075] In encoding each image block of each frame of the surveillance video, this embodiment uses H.265 encoding. Therefore, the t-th frame of the surveillance video... The adaptive QP value for encoding image blocks is defined as:

[0076] In the formula, These are the maximum and minimum values ​​of QP, which control the coding strength. The higher the QP value, the better the compression effect, but the higher the loss. When At higher levels, The smaller the value, the lower the degree of encoding compression for image blocks with high content complexity, ensuring the quality of the monitoring image. Conversely, a higher value indicates the use of higher encoding compression parameters to improve transmission efficiency. The size can be set by the implementer according to the implementation scenario, without special restrictions. In this embodiment... The values ​​are 40 and 15 respectively. The flowchart for determining the adaptive QP value during image block encoding is as follows: Figure 2as shown.

[0077] In step S005, the encoded frames of the monitoring video are exchanged securely.

[0078] The obtained monitoring video is encoded using the H.265 encoding method with the modified QP value to obtain a video data sequence, and then the video data sequence is encrypted using the AES encryption algorithm to ensure the security of the data. Both the H.265 encoding and the AES encryption algorithm are known technologies, and the specific process is not described in detail.

[0079] The video data sequence is then encapsulated into a network extraction layer unit, packaged into an MP4 format, and transmitted through an HTTP protocol. The receiving end decrypts and decodes the transmitted video data sequence after receiving it, and finally completes the low-latency secure exchange of the monitoring video.

[0080] It should be noted that the above-mentioned embodiments of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments. The above describes specific embodiments of the present application. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or may be advantageous.

[0081] Each embodiment in the present specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0082] The above-described embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; modifying the technical solutions described in the above embodiments, or equivalently replacing some technical features, does not change the essence of the corresponding technical solutions out of the scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A low-latency video security exchange method suitable for a video surveillance center, characterized in that, The method comprises the following steps: Collecting each frame of monitoring video, and uniformly dividing each frame of monitoring video into blocks; Analyzing the gray level difference between pixel points in corresponding position image blocks in each frame of monitoring video and its adjacent frame of monitoring video, and the motion change information of each frame of monitoring video relative to its adjacent frame of monitoring video, to determine the spatial domain complexity of each image block of each frame of monitoring video; Performing frequency domain transformation on each frame of monitoring video, and determining the frequency domain complexity of each image block of each frame of monitoring video through the amplitude difference between frequency domain points in corresponding position image blocks in each frame of monitoring video and its adjacent frame of monitoring video, and the linear change degree of the phase difference between frequency domain points in corresponding position image blocks; Combining the spatial domain complexity and the frequency domain complexity to determine the comprehensive content complexity of each image block of each frame of monitoring video, so as to correct the quantization parameter when video encoding is performed on each image block of each frame of monitoring video; Exchanging the encoded each frame of monitoring video safely; The determination of the comprehensive content complexity of each image block of each frame of monitoring video comprises: If the normalized value of the spatial domain complexity of any image block of each frame of monitoring video is less than or equal to a preset threshold value, the comprehensive content complexity of the any image block is the normalized value of the spatial domain complexity of the any image block; otherwise, the comprehensive content complexity of the any image block is the product of the normalized value of the spatial domain complexity and the normalized value of the frequency domain complexity of the any image block; The determination of the spatial domain complexity of each image block of each frame of monitoring video comprises: Calculating the mean value of the gray level difference between all pixel points in corresponding position image blocks in each frame of monitoring video and its previous frame of monitoring video; and determining the optical flow vector of all feature points in each image block of each frame of monitoring video by using an optical flow method; Based on the numerical distribution of the modulus of the optical flow vector, determining the displacement change degree of each image block of each frame of monitoring video; The spatial domain complexity is positively correlated with the mean value and the displacement change degree; The determination of the frequency domain complexity of each image block of each frame of monitoring video comprises: Calculating the fusion result of the amplitude difference between all frequency domain points in corresponding position image blocks in each frame of monitoring video and its previous frame of monitoring video, denoted as a first fusion value, and calculating the fusion result of the square of the amplitude of all frequency domain points in corresponding position image blocks in the previous frame of monitoring video of each frame of monitoring video, denoted as a second fusion value; Based on the first fusion value and the second fusion value, determining the amplitude difference degree of each image block of each frame of monitoring video; Calculating the phase difference between all frequency domain points in corresponding position image blocks in each frame of monitoring video and its previous frame of monitoring video, and performing linear fitting on the phase difference to obtain a residual term of linear fitting, calculating the fusion result of the square of all residual terms between corresponding position image blocks in each frame of monitoring video and its previous frame of monitoring video, denoted as a third fusion value; The frequency domain complexity is positively correlated with the amplitude difference degree and the third fusion value; The further determination of the amplitude difference degree comprises: Calculating the sum of the second fusion value and a preset value greater than 0, and the amplitude difference degree is the ratio of the first fusion value to the sum.

2. The low-latency video security exchange method for a video monitoring center according to claim 1, wherein, The displacement variation degree is a fusion result of a mean value and a discrete degree of a modulus of an optical flow vector of all feature points in each image block of each frame of the monitoring video.

3. The low-latency video security exchange method for a video monitoring center according to claim 1, wherein, The spatial domain complexity is a product of the mean value and the displacement variation degree.

4. The low-latency video security exchange method for a video monitoring center according to claim 1, wherein, The frequency domain complexity is a product of the amplitude difference degree and the third fusion value.

5. The low-latency video security exchange method for a video monitoring center according to claim 1, wherein, The modification is a quantization parameter when each image block of each frame of the monitoring video is video encoded, comprising: A maximum value and a minimum value of a preset quantization parameter, calculating a difference value between the maximum value and the minimum value, combining a comprehensive content complexity of each image block of each frame of the monitoring video, determining a modified quantization parameter when each image block of each frame of the monitoring video is video encoded.

6. The low-latency video security exchange method for a video monitoring center according to claim 5, wherein, A multiplication result of the comprehensive content complexity of each image block of each frame of the monitoring video and the difference value, the modified quantization parameter is a difference value between the maximum value and the multiplication result.

Images