Live stream intelligent switching method based on shot analysis

Abstract

The invention discloses an intelligent live stream switching method based on shot analysis. The method comprises the following steps: S1, acquiring live broadcast signals and live broadcast pictures of all cameras; s2, performing signal detection on the obtained live broadcast picture; s3, analyzing image features and continuity in the live broadcast picture; and S4, carrying out impression analysis on the live broadcast signal and the live broadcast picture, and automatically switching the live broadcast signal to an analysis machine position for live broadcast. According to the method, intelligent analysis can be carried out on pictures of all input camera positions, the live broadcast lens with the most ornamental value is accurately screened out through recommendation of the decision-making tree, the operation of switching broadcast is automatically completed, the problem that the optimal ornamental value of the lens in live broadcast is difficult to guarantee is effectively solved, the experience feeling of a user in the watching process is enhanced, and the user experience is improved. And meanwhile, the live broadcast is more interesting.

Description

Technical Field

[0001] The present invention relates to the technical field of image intelligent analysis, and in particular to a live stream intelligent switching method based on lens analysis. Background Art

[0002] With the rapid development of 5G Internet, the public's viewing preferences have also changed. In this context, live broadcasting has gradually gained attention and been widely used in many scenarios such as cultural tourism, education, and events. Unlike real-time live broadcasting, live broadcasting has a relatively large tolerance for delays. This feature gives live broadcasting technology the opportunity to create a better viewing experience and service for users.

[0003] sRGB defines the colors of the three primary colors red, green, and blue as the maximum value of the color when the other two colors are zero. In the CIE xy color coordinate system, red is located at [0.6400, 0.3300], green is at [0.3000, 0.6000], blue is at [0.1500, 0.0600], and white is D65 at [0.3127, 0.3290]. For any RGB color space, non-negative R, G, B cannot be represented outside the triangle defined by the primary colors, which is just within the color gamut of the human eye. sRGB also defines a nonlinear conversion between the intensity of the primary colors and the actual stored values. This curve is similar to the gamma response of a CRT monitor. Reproducing this curve is more important than the correct display of sRGB images on the monitor. This nonlinear conversion means that sRGB is very efficient in making the integer values ​​in the image file represent the colors that the human eye can distinguish. sRGB is not used by professionals in the printing industry because its color gamut is not large enough, especially the blue-green color gamut cannot represent all the colors in CMYK printing.

[0004] The existing live broadcast production method is to manually program the camera signals at different locations and broadcast them live to users. Although this method satisfies the automatic switching of cameras in live broadcast scenarios, since the programs are manually pre-programmed, the live broadcast picture cannot be guaranteed to be the best live shot of all camera locations in the current time period. The manual program programming method makes it difficult for users to appreciate the most entertaining program content, which to a certain extent affects the user's viewing experience and the quality improvement of the live broadcast. Summary of the invention

[0005] In order to solve the above problems, the present invention provides a live intelligent switching method based on shot analysis, comprising the following steps: S1. acquiring live broadcast signals and live broadcast images of all camera positions; S2. performing signal detection on the acquired live broadcast images; S3. analyzing the image features and continuity in the live broadcast images; S4. performing a perceptual analysis on the live broadcast signals and live broadcast images, and automatically switching the live broadcast signals to the analysis camera position for live broadcast.

[0006] Furthermore, the S2 step specifically includes the following sub-steps: S21. Setting an integrity threshold range to perform integrity detection on the image frame; S22. Performing clarity analysis on changes in pixel grayscale values ​​in the image frame; S23. Comparing the color value in the input image frame with sRGB, and calculating the color deviation value; S24. Detecting specific obstructions in the image frame, wherein the specific obstructions specifically include: stains on the camera and foreign objects blocking the screen.

[0007] Furthermore, the step S21 specifically includes: dividing the previous and subsequent continuous image frames into a number of block data, and sequentially comparing the RGB pixel values ​​of adjacent corresponding blocks, calculating the standard deviation of the pixel value of each corresponding block, and comparing and judging it with the integrity threshold range.

[0008] Furthermore, the S3 step specifically includes the following sub-steps: S31. Extracting feature data from the image frame, and normalizing the extracted feature data after arranging and packaging; S32. Performing continuity analysis on the image feature data based on the timestamp, and calculating the timestamp difference of the image feature data; S33. Making a judgment by comparing the timestamp difference of adjacent image feature data and combining it with the preset minimum broadcast length requirement.

[0009] Furthermore, the image feature data specifically includes: specific objects, whether the camera is moving, scene semantics, weather conditions, specific animals, and natural environment.

[0010] Furthermore, the step S33 specifically includes: when the current image feature sequence meets the minimum broadcast length requirement, executing step S4; when the current image feature sequence does not meet the minimum broadcast length requirement, keeping the current output signal unchanged and returning the data to step S2.

[0011] The present invention provides a live streaming intelligent switching method based on shot analysis, which has the following beneficial effects: The present invention abandons the traditional manual editing mode, and can carry out intelligent analysis of the images of all input cameras. Through the recommendation of the decision tree, it can accurately screen out the most ornamental live shots and automatically complete the operation of switching broadcasts, effectively solving the problem of difficulty in ensuring the best ornamental quality of shots in live broadcasts, enhancing the user experience during the viewing process, and making live broadcasts more interesting. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0013] Figure 1 The present invention provides a flow chart of the method. DETAILED DESCRIPTION

[0014] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0015] The following is a detailed description of the implementation method of the present invention in conjunction with the accompanying drawings. Only some embodiments are described, not all embodiments. For the purpose of clarity, representations and descriptions that are not related to the present invention are omitted in the drawings and descriptions.

[0016] In order to have a clearer understanding of the technical features, purposes and beneficial effects of the present invention, the technical solution of the present invention is now described in detail below. Obviously, the implementation cases described are part of the embodiments of the present invention, not all of the embodiments, and cannot be understood as limiting the scope of the implementation of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the protection scope of the present invention.

[0017] The present invention provides a live intelligent switching method based on shot analysis, such as Figure 1 As shown, the following steps are included: S1. Obtain the live broadcast signals and live broadcast images of all camera positions.

[0018] S2. Perform signal detection on the acquired live broadcast image: S21. Set the integrity threshold range and perform integrity detection on the image frames: divide the previous and subsequent continuous image frames into several block data, and compare the RGB pixel values ​​of adjacent corresponding blocks in turn, calculate the standard deviation of the pixel value of each corresponding block, and compare and judge it with the integrity threshold range.

[0019] S22. Perform clarity analysis on the change of pixel grayscale value in the image frame: Analyze the image clarity by changing the pixel grayscale value in the image, and compare the gradient value of the image frame with the preset clarity. The image whose standard deviation value meets the threshold is sent to the color restoration detection link.

[0020] S23. Compare the color values ​​in the input image frame with sRGB and calculate the color deviation value: For the pixel color values ​​in the input image frame, compare them with the SRGB color model, calculate the color deviation value of each pixel, and count the average color difference of the entire image. The image whose standard deviation value meets the threshold is sent to the frame occlusion detection link.

[0021] S24. Detect specific obstructions in the image frame, wherein the specific obstructions specifically include: stains on the camera, foreign objects blocking the image. Through target detection and measurement, specific obstructions (such as stains on the camera lens, foreign objects blocking the image, etc.) are detected, and images whose standard deviation values ​​meet the threshold are sent to the image feature analysis module.

[0022] S3. Analyze the image features and continuity in the live broadcast image.

[0023] Step S3 specifically includes the following sub-steps: S31. Extracting feature data from the image frame, organizing and packaging the extracted feature data, and then normalizing it; S32. Performing continuity analysis on the image feature data based on the timestamp, and calculating the timestamp difference of the image feature data; S33. Making a judgment by comparing the timestamp difference of adjacent image feature data and combining it with the preset minimum broadcast length requirement.

[0024] Image feature data specifically includes: specific objects (such as swings), whether the camera is moving, scene semantics (such as natural sentences such as jumping and running), weather conditions, specific animals (such as pandas), natural environment and other dimensional features.

[0025] Step S33 specifically includes: when the current image feature sequence meets the minimum broadcast length requirement, executing step S4; when the current image feature sequence does not meet the minimum broadcast length requirement, keeping the current output signal unchanged and returning the data to step S2.

[0026] S4. Analyze the visual quality of the input signal. By analyzing the visual characteristics of the live images of each input camera position, ensure that the output image is the most in line with the requirements of live broadcast among all current camera positions, and automatically switch the live broadcast signal to the analysis camera position for live broadcast.

[0027] It also includes a live stream intelligent switching system for implementing the method, including a signal frame extraction module, a signal detection module, an image feature analysis module, and a continuity analysis module.

[0028] The signal frame extraction module is used to perform interval frame extraction on the live broadcast signals of all input cameras according to the preset value. This module is responsible for performing interval frame extraction on the live broadcast signals of all input cameras according to the preset value. After extracting each frame of the image, it will accurately mark the corresponding timestamp for it. The timestamp is recorded in a relative time representation based on the start time of the live broadcast.

[0029] The signal detection module is used to detect the image signal and its clarity. The signal detection module specifically includes: image integrity detection, clarity detection, frame occlusion detection, and color restoration detection.

[0030] The image feature analysis module is used to extract and analyze features of image data. The image feature analysis module specifically includes: object recognition, semantic recognition, scene movement recognition, weather recognition, animal recognition, and natural environment recognition.

[0031] The continuity analysis module is used to perform continuity analysis on image feature data based on timestamps.

[0032] The present invention abandons the traditional manual editing mode, and can carry out intelligent analysis of the images of all input cameras. Through the recommendation of the decision tree, it can accurately screen out the most ornamental live shots and automatically complete the operation of switching broadcasts, effectively solving the problem of difficulty in ensuring the best ornamental quality of shots in live broadcasts, enhancing the user experience during the viewing process, and making live broadcasts more interesting.

[0033] The above is only a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art shall not deviate from the spirit and scope of the present invention, and shall be within the scope of protection of the claims attached to the present invention.

Claims

1. A method for intelligent switching of live streaming based on lens analysis, characterized in that: The following steps are involved: S1. Obtain the live broadcast signal and live broadcast screen of all cameras; S2. Perform signal detection on the acquired live screen; S3. Analyze the image features and continuity in the live broadcast screen; S4. Perform a visual analysis on the live broadcast signal and live broadcast picture, and automatically switch the live broadcast signal to the analysis camera position for live broadcast.

2. The method for intelligent switching of live streams based on lens analysis according to claim 1, characterized in that: The S2 step specifically includes the following sub-steps: S21. Set the integrity threshold range and perform integrity detection on the image frame; S22. Perform clarity analysis on the grayscale value changes of pixels in the image frame; S23. Compare the color value in the input image frame with sRGB and calculate the color deviation value; S24. Detect specific obstructions in the image frame, wherein the specific obstructions specifically include: stains on the camera and foreign objects blocking the image.

3. The method for intelligent switching of live streams based on lens analysis according to claim 2 is characterized in that: The step S21 specifically includes: dividing the previous and subsequent continuous image frames into a number of block data, and sequentially comparing the RGB pixel values ​​of adjacent corresponding blocks, calculating the standard deviation of the pixel value of each corresponding block, and comparing and judging it with the integrity threshold range.

4. The method for intelligent switching of live streams based on lens analysis according to claim 1, characterized in that: The S3 step specifically includes the following sub-steps: S31. Extracting feature data from the image frame, organizing and packaging the extracted feature data, and then normalizing them; S32. Performing continuity analysis on the image feature data based on the timestamp, and calculating the timestamp difference of the image feature data; S33. Making a judgment by comparing the timestamp difference of adjacent image feature data and combining it with the preset minimum broadcast length requirement.

5. The method for intelligent switching of live streams based on lens analysis according to claim 4 is characterized in that: The image feature data specifically includes: specific objects, whether the camera is moving, scene semantics, weather conditions, specific animals, and natural environment.

6. The method for intelligent switching of live streams based on lens analysis according to claim 4 is characterized in that: The step S33 specifically includes: when the current image feature sequence meets the minimum broadcast length requirement, executing step S4; when the current image feature sequence does not meet the minimum broadcast length requirement, keeping the current output signal unchanged and returning the data to step S2.

Images