Video image quality diagnosis and analysis detection system and detection method

Abstract

The present invention relates to the field of video image processing technology, and specifically discloses a video image quality diagnosis and analysis detection system and detection method. By extracting images from an input video stream frame by frame and performing preprocessing, the key feature areas in the image are identified based on edge detection and color segmentation algorithms, and consistency features, frame rate stability features, and spatial distribution density features are extracted. By calculating and analyzing these features, a consistency feature index, a frame rate stability feature index, and a spatial distribution density feature index are generated. Finally, the image quality score of each key feature area is obtained by comprehensive processing. Through this process, the quality change of the video image can be accurately evaluated to ensure that it meets the preset quality standards. The present invention also has an adaptive adjustment capability, and can dynamically adjust the parameters of image quality detection according to real-time monitoring results, such as optimizing the frame rate calculation threshold and improving the feature extraction accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of video image processing, and in particular to a video image quality diagnosis and analysis detection system and a detection method. Background Art

[0002] With the continuous development and widespread application of digital video technology, video image quality plays a vital role in many fields, especially in industries such as surveillance, medical imaging, video communications, and media production, where the requirements for video quality are becoming increasingly stringent. Video quality assessment usually involves image clarity, color accuracy, frame rate stability, and the overall performance of the video stream, which directly affect the user experience and the accuracy of subsequent data analysis. However, in a dynamic and complex environment, video quality may fluctuate dramatically due to factors such as network bandwidth, lighting changes, and video encoding compression. Therefore, timely and accurate detection and evaluation of video image quality has become a key technical issue in improving the effectiveness of video applications.

[0003] The prior art has the following deficiencies: Most of the existing image quality detection methods rely on traditional static analysis methods, such as conventional features such as image clarity, contrast, and noise. However, these methods often have significant deficiencies when facing dynamic video streams. On the one hand, traditional quality assessment methods cannot respond to quality fluctuations in video streams in real time, especially when the frame rate is unstable, edge information is lost, or the quality is degraded due to video compression, and cannot effectively make adaptive adjustments. On the other hand, when dealing with video quality in complex scenes, the existing technology often cannot combine multi-dimensional information such as the spatial distribution density, consistency characteristics, and frame rate stability of the image, which limits the accuracy of the assessment results and makes it difficult to meet the requirements of high-precision and high-real-time video quality control. Therefore, the existing technology has great limitations in the quality assessment of dynamic video streams, and there is an urgent need for an innovative solution that can evaluate video quality in real time and accurately and has adaptive adjustment capabilities to cope with the changing monitoring and analysis environment.

[0004] Therefore, the present application provides a video image quality diagnosis and analysis detection system and detection method to meet the needs. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a video image quality diagnosis and analysis detection system and a detection method to solve the existing problem of dynamic video stream quality evaluation.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: A video image quality diagnosis and analysis detection method comprises the following steps: S1: Extract images frame by frame from the input video stream and preprocess each frame; S2: Based on the preprocessed image, edge detection and color segmentation algorithms are used to identify key feature areas in the video, and key features are extracted. The key features include: consistency features, frame rate stability features, and spatial distribution density features; S3: analyzing the extracted consistency features, frame rate stability features, and spatial distribution density features, and calculating the consistency feature index, frame rate stability feature index, and spatial distribution density feature index respectively according to the analysis results; S4: performing comprehensive calculation processing on the calculated consistency feature index, frame rate stability feature index and spatial distribution density feature index to obtain an image quality score for each key feature area, which is used to evaluate the image quality of each key feature area; S5: According to the evaluation results, each key feature area is divided into a quality qualified area and an unqualified quality area. In combination with real-time monitoring, the quality change of the video stream is evaluated in real time, and the detection method is dynamically adjusted according to the current detection results.

[0007] Optionally, as a further solution of the present invention: the method of identifying key feature areas in the video based on the preprocessed image using edge detection and color segmentation algorithms specifically includes: Extract images frame by frame from the input video stream and preprocess each frame; Perform edge detection on the image and extract edge information in the image; Identify color areas in images based on color space conversion and threshold segmentation methods; The edge information is logically combined with the color segmentation results to identify the key feature areas in the image.

[0008] As a further solution of the present invention: the extraction process of the key features is: The extraction of consistent features includes: calculating the similarity between image frames, using image matching method, analyzing the consistency between different frames through mean square error, and extracting consistent features; Extraction of the frame rate stability feature includes: calculating the change amount between adjacent frames, calculating the change difference between adjacent frames through pixel difference measurement, and obtaining the frame rate stability feature; The extraction of spatial distribution density features includes: according to the distribution of edges in the image, using an edge detection algorithm to extract edge information of the image, and calculating the density distribution of the edge to obtain the spatial distribution density features.

[0009] As a further solution of the present invention: the analysis of the extracted consistency features, frame rate stability features and spatial distribution density features specifically includes: Obtain the consistency features of the key feature area, analyze the consistency abnormality between different frames of the video image, and calculate the consistency feature index based on the analysis results; Obtain the frame rate stability characteristics of the key feature area, analyze the degree of change between adjacent frames, and calculate the frame rate stability characteristic index according to the degree of frame rate fluctuation; The spatial distribution density characteristics of the key feature area are obtained, the complexity of the spatial information of the key feature area is analyzed, and the spatial distribution density characteristic index is calculated according to the size distribution of the edge density.

[0010] As a further solution of the present invention: the process of obtaining the consistency feature index is: Decomposing the video signal of the key feature area into multiple frames, and performing standardization processing on each frame image to obtain preprocessed image data; Using an autoencoder to train the preprocessed image data to obtain a low-dimensional feature representation of each frame, and calculating a reconstruction error of each frame of the image; Calculate the reconstruction error difference between two adjacent frames to obtain the consistency score between each frame and the previous frame. The calculation expression is: ; in, Indicates Frame and The consistency score between frames, Indicates The reconstruction error of the frame, Indicates the frame number; The consistency feature index is calculated by taking a weighted average of the consistency scores between all frames. The calculation expression is: ; In the formula, Represents the total number of video frames, Indicates The weight of the frame, is the consistency characteristic index.

[0011] As a further solution of the present invention: the process of obtaining the frame rate stability characteristic index is: Decompose the video signal into multiple frames and calculate the time interval between each frame; Initial frame rate estimate ,in Indicates the inter-frame time difference of the first frame, the initial error covariance , the state transfer matrix , the measurement matrix , process noise covariance ; Perform the recursive operation of Kalman filtering, including prediction and update steps: Among them, the calculation expression of the prediction step is: ; ; in, Indicates The frame rate prediction value of the frame, Represents the state transfer matrix The transposed matrix of represents the transpose operation, represents the error covariance of the prediction, Indicates the frame number; Among them, the calculation expression of the update step is: ; ; ; in, represents the Kalman gain, Indicates The updated frame rate estimate for the frame, represents the updated error covariance, Indicates The frame time interval, represents the identity matrix; Calculate the residual of each frame, the calculation expression is: ; in, Indicates The residual of the frame, Indicates The frame rate estimate of the frame; By calculating the standard deviation of all frame residuals, the frame rate stability characteristic index is obtained, and the calculation expression is: ; In the formula, Represents the frame rate stability characteristic index, Indicates the total number of video frames.

[0012] As a further solution of the present invention: the process of obtaining the spatial distribution density characteristic index is: Input Image ,in, Represents the two-dimensional coordinates of pixels in an image; The image is decomposed into four sub-bands using a two-dimensional discrete wavelet transform: low-frequency sub-band , horizontal high frequency sub-band , vertical high frequency sub-band and diagonal high frequency subbands ; For each subband, calculate the energy value, the calculation expression is: ; ; ; ; in, represents the energy value of the low-frequency subband, represents the energy value of the horizontal high-frequency subband, represents the energy value of the vertical high-frequency sub-band, represents the energy value of the diagonal high-frequency subband, Represents the row and column coordinates of the image; By calculating the ratio of the energy of the high-frequency sub-band to the energy of the low-frequency sub-band, the edge density of the image is obtained. The calculation expression is: ; in, Represents the edge density of the image; According to the edge density and the energy proportion of each sub-band, the spatial distribution density characteristic index is calculated, and the calculation expression is: ; in, Represents the spatial distribution density characteristic index, represents the energy of each subband, Indicates the number of subbands.

[0013] As a further solution of the present invention: the calculated consistency feature index, frame rate stability feature index and spatial distribution density feature index are comprehensively calculated and processed to obtain an image quality score for each key feature area, specifically including: Obtain the consistency feature index, frame rate stability feature index and spatial distribution density feature index of the key feature area, perform normalized comprehensive calculation processing on the consistency feature index, frame rate stability feature index and spatial distribution density feature index, and calculate the image quality score; The calculation expression of the image quality score is: ; In the formula, represents the image quality score, , and is the preset scale factor, and , and are greater than 0, represents the consistency feature index, Represents the frame rate stability characteristic index, Represents the spatial distribution density characteristic index.

[0014] As a further solution of the present invention: the method of dividing each key feature area into a quality qualified area and a quality unqualified area according to the evaluation results specifically includes: Calculate the image quality score of each key feature area respectively, and compare the image quality score of each key feature area with a preset threshold; Determine whether the image quality score of each key feature area is greater than or equal to the preset threshold. If so, it means that the video image quality of the corresponding area is qualified and is recorded as a qualified quality area. If not, it means that the video image quality of the corresponding area is unqualified and is recorded as an unqualified quality area.

[0015] A video image quality diagnosis and analysis detection system, comprising: A data acquisition and preprocessing module, wherein the data acquisition and preprocessing module extracts images frame by frame from the input video stream and preprocesses each frame; A feature recognition and extraction module, which uses edge detection and color segmentation algorithms to identify key feature areas in the video based on the preprocessed image and extract key features, including consistency features, frame rate stability features, and spatial distribution density features; A feature index calculation module, wherein the feature index calculation module analyzes the extracted consistency feature, frame rate stability feature, and spatial distribution density feature, and calculates a consistency feature index, a frame rate stability feature index, and a spatial distribution density feature index according to the analysis results; A comprehensive evaluation module, which performs comprehensive calculation processing on the calculated consistency feature index, frame rate stability feature index and spatial distribution density feature index to obtain an image quality score for each key feature area, which is used to evaluate the image quality of each key feature area; An image quality division module, wherein the image quality division module divides each key feature area into a qualified quality area and an unqualified quality area according to the evaluation result; Dynamically adjust the detection module, combined with real-time monitoring, to evaluate the quality changes of the video stream in real time, and dynamically adjust the detection method according to the current detection results.

[0016] Compared with the prior art, the present invention has at least the following beneficial effects: (1) In the above scheme, by setting up frame-by-frame extraction of images in the video stream and combining advanced edge detection and color segmentation algorithms, the key feature areas in the video are accurately identified, ensuring that each frame of the image can accurately capture important quality indicators. By analyzing consistency features, frame rate stability features and spatial distribution density features, the present invention can deeply evaluate the quality of video images and calculate the corresponding quality scores. In this process, the close combination of edge information and color segmentation results makes the identification of key feature areas more accurate, can effectively eliminate noise interference, and improve the accuracy of feature extraction. Based on these analyses, the system can evaluate in real time whether the video quality meets the preset standards, and provide an adaptive adjustment mechanism for dynamically changing video streams, quickly respond to quality fluctuations or anomalies, adjust detection strategies and algorithms, and thus optimize the effect of video quality control. This not only improves the accuracy and flexibility of detection, but also ensures that video quality evaluation can be carried out in real time and stably in a complex and changeable video stream environment, minimizing the risk of misjudgment and missed detection.

[0017] (2) In the above scheme, the parameters of image quality detection are flexibly adjusted by setting according to the real-time monitoring results and quality assessment. The system can dynamically adjust the threshold of frame rate calculation, optimize the accuracy of feature extraction, and ensure that the best detection performance is always maintained in different scenarios. This flexibility is particularly suitable for complex environments facing frequent quality fluctuations. When quality fluctuations occur, the system can automatically identify and adjust the detection strategy to ensure high-precision and high-stability image quality assessment. This mechanism significantly improves the robustness of the system under variable video stream conditions, especially in high-demand monitoring, real-time analysis and large-scale video processing applications, ensuring the continuous stability and reliability of video quality and meeting the quality control requirements in various dynamic environments. In addition, the automated adjustment process reduces the need for manual intervention and reduces the complexity of operation, providing a high degree of intelligence and autonomy for the intelligent video quality monitoring system, greatly improving the efficiency and accuracy of video quality monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings, which are incorporated herein and constitute a part of the specification, illustrate embodiments of the invention and, together with the description, further serve to explain the principles of the invention and to enable those skilled in the relevant art to make and use the invention.

[0019] Figure 1 A flowchart of a video image quality diagnosis and analysis detection method; Figure 2 The present invention is a flowchart of a video image quality diagnosis and analysis detection system. DETAILED DESCRIPTION

[0020] The following is a detailed description of a video image quality diagnosis and analysis detection method provided by the present invention in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0021] See also Figure 1 As shown, the present invention is a video image quality diagnosis and analysis detection method, comprising the following steps: S1: Extract images frame by frame from the input video stream and preprocess each frame; S2: Based on the preprocessed image, edge detection and color segmentation algorithms are used to identify key feature areas in the video, and key features are extracted. The key features include: consistency features, frame rate stability features, and spatial distribution density features; S3: analyzing the extracted consistency features, frame rate stability features, and spatial distribution density features, and calculating the consistency feature index, frame rate stability feature index, and spatial distribution density feature index respectively according to the analysis results; S4: performing comprehensive calculation processing on the calculated consistency feature index, frame rate stability feature index and spatial distribution density feature index to obtain an image quality score for each key feature area, which is used to evaluate the image quality of each key feature area; S5: According to the evaluation results, each key feature area is divided into a quality qualified area and an unqualified quality area. In combination with real-time monitoring, the quality change of the video stream is evaluated in real time, and the detection method is dynamically adjusted according to the current detection results.

[0022] In S1, images are extracted frame by frame from the input video stream and each frame is preprocessed, including: Read image data frame by frame from the input video stream to ensure that the resolution and frame rate parameters of each frame remain consistent during extraction; According to the frame rate of the video stream, frames are extracted regularly to ensure that each extracted frame is consistent with the playback speed of the video and to prevent excessive or insufficient frame extraction from adversely affecting subsequent processing; Convert each frame of color image into grayscale image and use Gaussian filtering to remove noise in the image, especially in video streams, where there is motion blur or transmission noise; After the above preprocessing, each frame is saved as processed image data for use in the subsequent key feature extraction step.

[0023] In S2, based on the preprocessed image, edge detection and color segmentation algorithms are used to identify key feature areas in the video, and key features are extracted. The key features include: consistency features, frame rate stability features, and spatial distribution density features, specifically including: Extract images frame by frame from the input video stream and preprocess each frame, including: extract each frame to ensure that its time interval matches the frame rate of the video stream to avoid frame loss or excessive frame extraction. Preprocess each frame, including grayscale, denoising, and image enhancement to ensure that the image quality meets the standards for edge detection and color region recognition; after preprocessing, pass each frame of the image as input to the subsequent edge detection and color region recognition steps.

[0024] Perform edge detection on the image and extract edge information in the image, including: The purpose of edge detection is to identify significant edge areas in the image, which are usually the boundaries of objects, structures or textures. The Canny edge detection algorithm uses Gaussian filtering to remove noise, calculates gradient amplitude and direction, and applies a double threshold strategy to extract edge pixels in the image, forming a binary edge map and identifying edge areas.

[0025] Based on color space conversion and threshold segmentation methods, color regions in the image are identified, including: Color space conversion helps extract color features under different lighting conditions. Based on the set color threshold, the target color region in the image is identified through color segmentation. This process groups colors by setting color ranges to extract important color regions in the image, corresponding to objects or scenes in the video.

[0026] The edge information is logically combined with the color segmentation results to identify the key feature areas in the image, including: combining edge and color information through logical operations to accurately identify the key feature areas in the image. These areas usually correspond to the objects of interest or key parts in the video, such as the edges of objects, the outlines of people, or significant color blocks. Through this process, representative key feature areas can be extracted for subsequent image quality analysis and evaluation.

[0027] The key feature extraction process is as follows: The extraction of consistent features includes: calculating the similarity between image frames, using image matching method, analyzing the consistency between different frames through mean square error, and extracting consistent features; Extraction of the frame rate stability feature includes: calculating the change amount between adjacent frames, calculating the change difference between adjacent frames through pixel difference measurement, and obtaining the frame rate stability feature; The extraction of spatial distribution density features includes: according to the distribution of edges in the image, using an edge detection algorithm to extract edge information of the image, and calculating the density distribution of the edge to obtain the spatial distribution density features.

[0028] In S3, the extracted consistency features, frame rate stability features, and spatial distribution density features are analyzed, and according to the analysis results, the consistency feature index, frame rate stability feature index, and spatial distribution density feature index are calculated respectively, including: Obtain the consistency features of the key feature area, analyze the consistency abnormality between different frames of the video image, and calculate the consistency feature index based on the analysis results; Obtain the frame rate stability characteristics of the key feature area, analyze the degree of change between adjacent frames, and calculate the frame rate stability characteristic index according to the degree of frame rate fluctuation; Obtain the spatial distribution density characteristics of the key feature area, analyze the complexity of the spatial information of the key feature area, and calculate the spatial distribution density characteristic index according to the size distribution of the edge density; The process of obtaining the consistency feature index is as follows: Decomposing the video signal of the key feature area into multiple frames, and performing standardization processing on each frame image to obtain preprocessed image data; Using an autoencoder to train the preprocessed image data to obtain a low-dimensional feature representation of each frame, and calculating a reconstruction error of each frame of the image; Calculate the reconstruction error difference between two adjacent frames to obtain the consistency score between each frame and the previous frame. The calculation expression is: ; in, Indicates Frame and The consistency score between frames, Indicates The reconstruction error of the frame, Indicates the frame number; The consistency feature index is calculated by taking a weighted average of the consistency scores between all frames. The calculation expression is: ; In the formula, Represents the total number of video frames, Indicates The weight of the frame, is the consistency characteristic index; The process of obtaining the frame rate stability characteristic index is as follows: Decompose the video signal into multiple frames and calculate the time interval between each frame; Initial frame rate estimate ,in Indicates the inter-frame time difference of the first frame, the initial error covariance , the state transfer matrix , the measurement matrix , process noise covariance ; Perform the recursive operation of Kalman filtering, including prediction and update steps: Among them, the calculation expression of the prediction step is: ; ; in, Indicates The frame rate prediction value of the frame, Represents the state transfer matrix The transposed matrix of represents the transpose operation, represents the error covariance of the prediction, Indicates the frame number; Among them, the calculation expression of the update step is: ; ; ; in, represents the Kalman gain, Indicates The updated frame rate estimate for the frame, represents the updated error covariance, Indicates The frame time interval, represents the identity matrix; Calculate the residual of each frame, the calculation expression is: ; in, Indicates The residual of the frame, Indicates The frame rate estimate of the frame; By calculating the standard deviation of all frame residuals, the frame rate stability characteristic index is obtained, and the calculation expression is: ; In the formula, Represents the frame rate stability characteristic index, Indicates the total number of video frames; The process of obtaining the spatial distribution density characteristic index is as follows: Input Image ,in, Represents the two-dimensional coordinates of pixels in an image; The image is decomposed into four sub-bands using a two-dimensional discrete wavelet transform: low-frequency sub-band , horizontal high frequency sub-band , vertical high frequency sub-band and diagonal high frequency subbands ; For each subband, calculate the energy value, the calculation expression is: ; ; ; ; in, represents the energy value of the low-frequency subband, represents the energy value of the horizontal high-frequency subband, represents the energy value of the vertical high-frequency sub-band, represents the energy value of the diagonal high-frequency subband, Represents the row and column coordinates of the image; By calculating the ratio of the energy of the high-frequency sub-band to the energy of the low-frequency sub-band, the edge density of the image is obtained. The calculation expression is: ; in, Represents the edge density of the image; According to the edge density and the energy proportion of each sub-band, the spatial distribution density characteristic index is calculated, and the calculation expression is: ; in, Represents the spatial distribution density characteristic index, represents the energy of each subband, Indicates the number of subbands.

[0029] In S4, the calculated consistency feature index, frame rate stability feature index and spatial distribution density feature index are comprehensively calculated and processed to obtain an image quality score for each key feature area, which is used to evaluate the image quality of each key feature area, specifically including: Obtain the consistency feature index, frame rate stability feature index and spatial distribution density feature index of the key feature area, perform normalized comprehensive calculation processing on the consistency feature index, frame rate stability feature index and spatial distribution density feature index, and calculate the image quality score; The calculation expression of the image quality score is: ; In the formula, represents the image quality score, , and is the preset scale factor, and , and are greater than 0, represents the consistency feature index, Represents the frame rate stability characteristic index, Represents the spatial distribution density characteristic index; It should be noted that the image quality score is used to evaluate whether the quality of the video image is qualified during quality detection, and the larger the image quality score of the video image, the better the quality of the corresponding video image.

[0030] In S5, each key feature area is divided into a qualified quality area and an unqualified quality area according to the evaluation results, and combined with real-time monitoring, the quality change of the video stream is evaluated in real time, and the detection method is dynamically adjusted according to the current detection results, including: Calculate the image quality score of each key feature area respectively, and compare the image quality score of each key feature area with a preset threshold; Determine whether the image quality score of each key feature area is greater than or equal to a preset threshold. If so, it means that the video image quality of the corresponding area is qualified and recorded as a qualified quality area. If not, it means that the video image quality of the corresponding area is unqualified and recorded as an unqualified quality area. The above-mentioned method is combined with real-time monitoring to evaluate the quality change of the video stream in real time, and dynamically adjust the detection method according to the current detection results, specifically including: Each frame is continuously extracted from the video stream and preprocessed. The real-time extraction process is adjusted according to the actual frame rate and resolution of the video stream to ensure the timeliness of video quality analysis.

[0031] Consistency features, frame rate stability features and spatial distribution density features are extracted for each frame, and key feature areas are acquired in real time by combining edge detection and color segmentation techniques.

[0032] The consistency feature index, frame rate stability feature index and spatial distribution density feature index are obtained through real-time calculation and normalized to calculate the image quality score of each frame.

[0033] Based on the comparison between the image quality score and the threshold, each key feature area is divided into a qualified quality area or an unqualified quality area in real time, and the quality assessment results are recorded to provide a basis for subsequent dynamic adjustments.

[0034] Combined with the real-time quality assessment results of each frame, the overall quality of the video stream is monitored and fed back to the quality control system for further adjustment of video detection strategies and methods.

[0035] According to the real-time monitoring results and quality assessment, the image quality detection parameters are dynamically adjusted, such as changing the threshold of frame rate calculation, the accuracy of feature extraction, etc. If the video stream quality is found to be degraded or the frame rate fluctuates, the detection strategy can be automatically adjusted to optimize the detection results.

[0036] According to the results of real-time analysis, dynamically adjust the extraction methods of consistency features, frame rate stability features, and spatial distribution density features. If areas repeatedly fail to meet quality standards, increase the detection frequency for these areas.

[0037] Dynamically adjust the preset threshold of the quality score based on the change of video stream quality. If a long period of quality degradation is detected, adjust the standard for dividing qualified and unqualified areas to ensure more accurate detection results.

[0038] See also Figure 2 As shown, a video image quality diagnosis and analysis detection system comprises: A data acquisition and preprocessing module, wherein the data acquisition and preprocessing module extracts images frame by frame from the input video stream and preprocesses each frame; A feature recognition and extraction module, which uses edge detection and color segmentation algorithms to identify key feature areas in the video based on the preprocessed image and extract key features, including consistency features, frame rate stability features, and spatial distribution density features; A feature index calculation module, wherein the feature index calculation module analyzes the extracted consistency feature, frame rate stability feature, and spatial distribution density feature, and calculates a consistency feature index, a frame rate stability feature index, and a spatial distribution density feature index according to the analysis results; A comprehensive evaluation module, which performs comprehensive calculation processing on the calculated consistency feature index, frame rate stability feature index and spatial distribution density feature index to obtain an image quality score for each key feature area, which is used to evaluate the image quality of each key feature area; An image quality division module, wherein the image quality division module divides each key feature area into a qualified quality area and an unqualified quality area according to the evaluation result; Dynamically adjust the detection module, combined with real-time monitoring, to evaluate the quality changes of the video stream in real time, and dynamically adjust the detection method according to the current detection results.

[0039] The working principle provided by the present invention is to extract the images in the video stream frame by frame and perform preprocessing, and combine edge detection and color segmentation algorithms to extract the key feature areas in the video image, including: consistency features, frame rate stability features and spatial distribution density features. The extracted features are analyzed, the consistency feature index, frame rate stability feature index and spatial distribution density feature index are calculated, and the image quality score of each key feature area is obtained by comprehensive calculation, which is used to evaluate the regional image quality. Based on the evaluation results, the qualified and unqualified quality areas are dynamically divided, and the video stream quality is evaluated in real time in combination with real-time monitoring, and the detection strategy and method are automatically adjusted. The present invention can not only efficiently and accurately evaluate the video quality, but also improve the sensitivity and accuracy of detection through dynamic adjustment, ensuring the real-time and reliability of video images in quality control.

[0040] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0041] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A video image quality diagnosis and analysis detection method, characterized in that: The following steps are involved: S1: Extract images frame by frame from the input video stream and preprocess each frame; S2: Based on the preprocessed image, edge detection and color segmentation algorithms are used to identify key feature areas in the video, and key features are extracted. The key features include: consistency features, frame rate stability features, and spatial distribution density features; S3: analyzing the extracted consistency features, frame rate stability features, and spatial distribution density features, and calculating the consistency feature index, frame rate stability feature index, and spatial distribution density feature index respectively according to the analysis results; S4: performing comprehensive calculation processing on the calculated consistency feature index, frame rate stability feature index and spatial distribution density feature index to obtain an image quality score for each key feature area, which is used to evaluate the image quality of each key feature area; S5: According to the evaluation results, each key feature area is divided into a quality qualified area and an unqualified quality area. In combination with real-time monitoring, the quality change of the video stream is evaluated in real time, and the detection method is dynamically adjusted according to the current detection results.

2. A video image quality diagnosis and analysis detection method according to claim 1, characterized in that: The method of using edge detection and color segmentation algorithms to identify key feature areas in the video based on the preprocessed image specifically includes: Extract images frame by frame from the input video stream and preprocess each frame; Perform edge detection on the image and extract edge information in the image; Identify color areas in images based on color space conversion and threshold segmentation methods; The edge information is logically combined with the color segmentation results to identify the key feature areas in the image.

3. The video image quality diagnosis and analysis detection method according to claim 1, characterized in that: The key feature extraction process is as follows: The extraction of consistent features includes: extracting consistent features by calculating the similarity between image frames, using image matching methods, and analyzing the consistency between different frames through mean square error; Extraction of the frame rate stability feature includes: calculating the change amount between adjacent frames, calculating the change difference between adjacent frames through pixel difference measurement, and obtaining the frame rate stability feature; The extraction of spatial distribution density features includes: according to the distribution of edges in the image, using an edge detection algorithm to extract edge information of the image, and calculating the density distribution of the edge to obtain the spatial distribution density features.

4. The video image quality diagnosis and analysis detection method according to claim 1, characterized in that: The analysis of the extracted consistency features, frame rate stability features and spatial distribution density features specifically includes: Obtain the consistency features of the key feature areas, analyze the consistency abnormality between different frames of the video image, and calculate the consistency feature index based on the analysis results; Obtain the frame rate stability characteristics of the key feature area, analyze the degree of change between adjacent frames, and calculate the frame rate stability characteristic index according to the degree of frame rate fluctuation; The spatial distribution density characteristics of the key feature area are obtained, the complexity of the spatial information of the key feature area is analyzed, and the spatial distribution density characteristic index is calculated according to the size distribution of the edge density.

5. A video image quality diagnosis and analysis detection method according to claim 4, characterized in that: The process of obtaining the consistency feature index is as follows: Decomposing the video signal of the key feature area into multiple frames, and performing standardization processing on each frame image to obtain preprocessed image data; Using an autoencoder to train the preprocessed image data to obtain a low-dimensional feature representation of each frame, and calculating a reconstruction error of each frame of the image; Calculate the reconstruction error difference between two adjacent frames to obtain the consistency score between each frame and the previous frame. The calculation expression is: ； in, Indicates Frame and The consistency score between frames, Indicates The reconstruction error of the frame, Indicates the frame number; The consistency feature index is calculated by taking a weighted average of the consistency scores between all frames. The calculation expression is: ； In the formula, Indicates the total number of video frames, Indicates The weight of the frame, is the consistency characteristic index.

6. A video image quality diagnosis and analysis detection method according to claim 4, characterized in that: The process of obtaining the frame rate stability characteristic index is as follows: Decompose the video signal into multiple frames and calculate the time interval between each frame; Initial frame rate estimate ,in Indicates the inter-frame time difference of the first frame, the initial error covariance , the state transfer matrix , the measurement matrix , process noise covariance ; Perform the recursive operation of Kalman filtering, including prediction and update steps: Among them, the calculation expression of the prediction step is: ； ； in, Indicates The frame rate prediction value of the frame, Represents the state transfer matrix The transposed matrix of represents the transpose operation, represents the error covariance of the prediction, Indicates the frame number; Among them, the calculation expression of the update step is: ； ； ； in, represents the Kalman gain, Indicates The updated frame rate estimate for the frame, represents the updated error covariance, Indicates The frame time interval, represents the identity matrix; Calculate the residual of each frame, the calculation expression is: ； in, Indicates The residual of the frame, Indicates The frame rate estimate of the frame; By calculating the standard deviation of all frame residuals, the frame rate stability characteristic index is obtained, and the calculation expression is: ； In the formula, Represents the frame rate stability characteristic index, Indicates the total number of video frames.

7. The video image quality diagnosis and analysis detection method according to claim 4, characterized in that: The process of obtaining the spatial distribution density characteristic index is as follows: Input Image ,in, Represents the two-dimensional coordinates of pixels in an image; The image is decomposed into four sub-bands using a two-dimensional discrete wavelet transform: low-frequency sub-band , horizontal high frequency sub-band , vertical high frequency sub-band and diagonal high frequency subbands ; For each subband, calculate the energy value, the calculation expression is: ； ； ； ； in, represents the energy value of the low-frequency subband, represents the energy value of the horizontal high-frequency subband, represents the energy value of the vertical high-frequency sub-band, represents the energy value of the diagonal high-frequency subband, Represents the row and column coordinates of the image; By calculating the ratio of the energy of the high-frequency sub-band to the energy of the low-frequency sub-band, the edge density of the image is obtained. The calculation expression is: ； in, Represents the edge density of the image; According to the edge density and the energy proportion of each sub-band, the spatial distribution density characteristic index is calculated, and the calculation expression is: ； in, Represents the spatial distribution density characteristic index, represents the energy of each subband, Indicates the number of subbands.

8. The video image quality diagnosis and analysis detection method according to claim 1, characterized in that: The calculated consistency feature index, frame rate stability feature index and spatial distribution density feature index are comprehensively calculated and processed to obtain an image quality score for each key feature area, specifically including: Obtain the consistency feature index, frame rate stability feature index and spatial distribution density feature index of the key feature area, perform normalized comprehensive calculation processing on the consistency feature index, frame rate stability feature index and spatial distribution density feature index, and calculate the image quality score; The calculation expression of the image quality score is: ； In the formula, represents the image quality score, , and is the preset scale factor, and , and are greater than 0, represents the consistency feature index, Represents the frame rate stability characteristic index, Represents the spatial distribution density characteristic index.

9. The video image quality diagnosis and analysis detection method according to claim 1, characterized in that: According to the evaluation results, each key feature area is divided into a quality qualified area and a quality unqualified area, specifically including: Calculate the image quality score of each key feature area respectively, and compare the image quality score of each key feature area with a preset threshold; Determine whether the image quality score of each key feature area is greater than or equal to the preset threshold. If so, it means that the video image quality of the corresponding area is qualified and is recorded as a qualified quality area. If not, it means that the video image quality of the corresponding area is unqualified and is recorded as an unqualified quality area.

10. A video image quality diagnosis and analysis detection system, characterized in that: A video image quality diagnosis and analysis detection method as claimed in any one of claims 1 to 9, comprising: A data acquisition and preprocessing module, wherein the data acquisition and preprocessing module extracts images frame by frame from the input video stream and preprocesses each frame; A feature recognition and extraction module, which uses edge detection and color segmentation algorithms to identify key feature areas in the video based on the preprocessed image and extract key features, including consistency features, frame rate stability features, and spatial distribution density features; A feature index calculation module, wherein the feature index calculation module analyzes the extracted consistency feature, frame rate stability feature, and spatial distribution density feature, and calculates a consistency feature index, a frame rate stability feature index, and a spatial distribution density feature index according to the analysis results; A comprehensive evaluation module, which performs comprehensive calculation processing on the calculated consistency feature index, frame rate stability feature index and spatial distribution density feature index to obtain an image quality score for each key feature area, which is used to evaluate the image quality of each key feature area; An image quality division module, wherein the image quality division module divides each key feature area into a qualified quality area and an unqualified quality area according to the evaluation result; Dynamically adjust the detection module, combined with real-time monitoring, to evaluate the quality changes of the video stream in real time, and dynamically adjust the detection method according to the current detection results.

Images