A management platform with audio and video local storage management function
By mapping the video into a three-dimensional space and clustering data points based on spatial distance and grayscale differences, the scanning traversal direction is determined, solving the problem of low continuity and repetition caused by fixed scanning traversal paths, and achieving more efficient video compression and storage management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-26
- Publication Date
- 2026-03-24
AI Technical Summary
In existing technologies, when video images are scanned and encoded using a fixed scanning path, the degree of continuous repetition is low, resulting in poor video storage management.
By mapping the video into a three-dimensional space, clustering is performed based on the spatial distance and grayscale differences between data points to determine the main, secondary, and final directions of the scanning traversal. The scanning traversal path is then adaptively adjusted to improve the degree of continuous repetition or similarity, thereby enhancing the encoding and compression effect.
It improves the compression and storage management of videos, saving storage space and reducing storage costs.
Smart Images
Figure CN121397241B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of audio and video management, and particularly relates to a management platform with audio and video local storage management function. BACKGROUND
[0002] A large amount of audio and video data will be generated in the fields of security monitoring, remote conference, medical imaging, etc., and in order to save storage space, etc., it is usually necessary to compress and store the audio and video data. In the prior art, a fixed scanning traversal path is usually used to scan and traverse the video images in the audio and video, the scanning sequence obtained is run-length coded to obtain compressed data, and the compressed data is stored and managed. However, the degree of continuous repetition of video images in different fields in different directions will be different, and run-length coding has a good compression effect on continuous repeated pixels. Therefore, the scanning and traversal of the images in the video according to the fixed scanning traversal path will result in a relatively low degree of continuous repetition of the obtained sequence, and when the degree of continuous repetition is low, the storage management effect of the video will be poor. Therefore, the compression and storage management effect of the video needs to be improved. SUMMARY
[0003] In order to solve the above problems, the present application provides a management platform with audio and video local storage management function, and the technical solution is as follows:
[0004] One embodiment of the present application provides a management platform with audio and video local storage management function, which comprises:
[0005] A video mapping module is configured to obtain a to-be-compressed video and map each frame image in the to-be-compressed video to a three-dimensional space to obtain data points in the three-dimensional space.
[0006] A scanning traversal direction obtaining module is configured to obtain a metric distance between the data points in the three-dimensional space according to the spatial distance and the gray difference between the data points in the three-dimensional space, cluster the data points in the three-dimensional space according to the metric distance between the data points to obtain each cluster, and obtain a main scanning traversal direction, a secondary scanning traversal direction and a terminal scanning traversal direction according to the number of neighborhood data points of the data points in each cluster, the edge length of the minimum circumscribed cube of each cluster and the proportion of the data points in the cluster in the corresponding minimum circumscribed cube.
[0007] The compression storage management module is configured to encode the data points in the three-dimensional space according to the main scanning traversal direction, the secondary scanning traversal direction and the end scanning traversal direction to obtain compressed data, and store and manage the compressed data.
[0008] Beneficial effects: The application comprises a video mapping module configured to obtain a video to be compressed and map each frame image in the video to be compressed to data points in a three-dimensional space; a scanning traversal direction obtaining module configured to obtain a metric distance between the data points in the three-dimensional space according to a spatial distance and a gray difference between the data points in the three-dimensional space, and cluster the data points in the three-dimensional space according to the metric distance between the data points in the three-dimensional space to obtain each cluster, and obtain the main scanning traversal direction, the secondary scanning traversal direction and the end scanning traversal direction according to a number of neighborhood data points of the data points in each cluster, a side length of a minimum circumscribed cube of each cluster and a proportion of intra-cluster data points in the corresponding minimum circumscribed cube; and a compression storage management module configured to encode the data points in the three-dimensional space according to the main scanning traversal direction, the secondary scanning traversal direction and the end scanning traversal direction to obtain compressed data, and store and manage the compressed data. The main scanning traversal direction, the secondary scanning traversal direction and the end scanning traversal direction determined according to the cluster distribution characteristics can make the sequence obtained by scanning traversal have a high degree of continuous repetition or similarity, thereby improving the compression storage management effect of the video. BRIEF DESCRIPTION OF DRAWINGS
[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, a brief introduction will be given to the drawings needed in the embodiments or the prior art description. Obviously, the drawings in the following description only show some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without any creative effort.
[0010] Figure 1 The structural block diagram of the management platform with the audio and video local storage management function. DETAILED DESCRIPTION
[0011] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art belong to the scope of protection of the embodiments of the present application.
[0012] 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 the present application belongs.
[0013] The embodiment provides a management platform with audio and video local storage management functions, and details are as follows.
[0014] As shown in the figure, the embodiment provides a management platform with audio and video local storage management functions, which comprises: Figure 1
[0015] A video mapping module 01 is configured to acquire a to-be-compressed video, and map each frame image in the to-be-compressed video into a three-dimensional space to obtain data points in the three-dimensional space.
[0016] Since the continuous repetition or continuous similarity degree of images in different videos in different directions is different, the way of scanning and traversing the images in the video according to the fixed scanning and traversing path can result in low continuous repetition or continuous similarity degree of pixels in the obtained scanning sequence, which leads to poor encoding and compression effect of the scanning sequence, and thus leads to poor storage management effect of the video. In order to improve the compression effect of the video and ensure the effect of the video storage management, the embodiment will adaptively adjust the scanning and traversing path based on the video characteristics in the subsequent process, so as to improve the continuous repetition or continuous similarity degree of pixel values in the obtained scanning sequence, and thus improve the effect of the encoding and compression. That is, when the continuous repetition or continuous similarity degree of pixels in the scanning sequence is higher, the effect of the run-length encoding of the difference sequence of the scanning sequence is better or the compression efficiency is better. In addition, in order to facilitate understanding, the subsequent process will be performed on any video that needs to be compressed for compression storage management, and the video that needs to be compressed is recorded as a to-be-compressed video. The images in the to-be-compressed video in the embodiment are gray-scale images, the images in the to-be-compressed video are two-dimensional images, and when the compression storage management of a certain video is performed, the corresponding video needs to be transmitted to the video local storage management function management platform terminal, and the subsequent process is also performed on the video local storage management function management platform terminal.
[0017] After obtaining the video to be compressed, a three-dimensional space is constructed, the horizontal axis of the three-dimensional space represents the horizontal pixel position in a single frame image, the vertical axis represents the vertical pixel position in a single frame image, and the vertical axis represents the sequence number of the image in the video to be compressed. Then, each frame image in the video to be compressed is sequentially mapped into the three-dimensional space according to the order in the video, to obtain data points in the three-dimensional space, that is, there are multiple data points in the mapped three-dimensional space. The number of data points in the three-dimensional space is the number of all pixel points in the video to be compressed. During the mapping, each frame image is sequentially inserted into a layer of the three-dimensional space according to the order of the images in the video to be compressed. Each layer of the mapped three-dimensional space horizontally is a two-dimensional plane image, and multiple layers of images are stacked in the vertical direction of the mapped three-dimensional space, and the number of stacked image layers is consistent with the total number of images in the video. The coordinates of any pixel point on any image and the sequence number of the image to which the pixel point belongs in the video are mapped into the three-dimensional space to obtain a data point. The gray value of the data point is the gray value of the pixel point, that is, each data point in the three-dimensional space corresponds to a pixel point or a gray value. The spatial distance between any two adjacent data points in the three-dimensional space is one unit length in the horizontal direction, the vertical direction or the vertical direction, that is, the spatial distance between any two adjacent data points in any coordinate axis direction in the three-dimensional space is one unit length.
[0018] The horizontal coordinate of the data point in the mapped three-dimensional space represents the position index in the horizontal direction of a single frame image. For example, if there are 800 pixel points in the width direction of a frame image, the horizontal coordinate is from 1 to 800. The horizontal coordinate can determine the column in which the corresponding data point is located in the horizontal direction of the image. The vertical coordinate of the data point in the mapped three-dimensional space represents the position index in the vertical direction of a single frame image. For example, if there are 600 pixel points in the height direction of a frame image, the vertical coordinate is from 1 to 600. The vertical coordinate can determine the row in which the corresponding data point is located in the vertical direction of the image. The vertical coordinate of the data point in the mapped three-dimensional space represents the sequence number of the video frame. For example, if the video to be compressed has a total of 120 frames, that is, 120 images, the vertical coordinate is from 1 to 120. It determines the frame number of the corresponding point in the video. For example, for any data point in the three-dimensional space, if the coordinates of the data point are (1, 2, 3), it means that the data point is mapped from the pixel point at the position of the first column and the second row in the third frame image in the video to be compressed. The gray value corresponding to the data point is the gray value of the pixel point at the position of the first column and the second row in the third frame image. Or it means that the data point is mapped from the pixel point with coordinates (1, 2) in the third frame image in the video to be compressed. It should be noted that the horizontal axis and the vertical axis of the two-dimensional space before image mapping are consistent with the horizontal axis and the vertical axis of the three-dimensional space.
[0019] Therefore, the embodiment completes the conversion mapping of the video to be compressed through the above process, and obtains data points in a three-dimensional space. The embodiment maps the video to be compressed into the three-dimensional space because there may be similar or identical regions not only on a single frame image but also between different frame images. Therefore, all images in the video are mapped into the same space, and the similarity or identity between frames is considered when determining the scanning traversal path subsequently, so as to filter a scanning traversal path with higher repetition, and further improve the compressible upper limit of the run-length coding or further improve the compression effect of the run-length coding.
[0020] The scanning traversal direction obtaining module 02 is configured to obtain a metric distance between the data points in the three-dimensional space according to the spatial distance and the gray difference between the data points in the three-dimensional space, and cluster the data points in the three-dimensional space according to the metric distance between the data points in the three-dimensional space to obtain each cluster. The scanning traversal main direction, the scanning traversal secondary direction, and the scanning traversal terminal direction are obtained according to the number of neighborhood data points of the data points in each cluster, the side length of the minimum circumscribed cube of each cluster, and the proportion of the data points in the corresponding minimum circumscribed cube.
[0021] After completing the conversion mapping of the video, the embodiment clusters the data points in the three-dimensional space. The purpose of clustering is to facilitate the determination of the scanning traversal direction. Before clustering, the metric distance between the data points is determined based on the spatial distance and the gray difference between the data points. The metric distance is the basis for clustering the data points. Therefore, in the embodiment, the metric distance between the data points in the three-dimensional space is obtained according to the spatial distance and the gray difference between the data points in the three-dimensional space.
[0022] For the data point a1 and the data point a2 in the three-dimensional space, the coordinate Euclidean distance between the data point a1 and the data point a2 is calculated, and the coordinate Euclidean distance between the data point a1 and the data point a2 is normalized. The result of the normalization is recorded as a first distance value. The coordinate Euclidean distance between the data point a1 and the data point a2 is the Euclidean distance between the three-dimensional coordinates of the data point a1 and the three-dimensional coordinates of the data point a2. The absolute value of the difference between the gray value corresponding to the data point a1 and the gray value corresponding to the data point a2 is calculated. The absolute value is normalized. The result of the normalization is recorded as a second distance value. If a data point is mapped from a pixel point on a frame image, the gray value of the data point is the gray value corresponding to the pixel point. The sum of the first distance value and the second distance value is calculated and used as the metric distance between the data point a1 and the data point a2. The calculation expression of the metric distance between the data point a1 and the data point a2 is:
[0023]
[0024] wherein, is a metric distance between data point a1 and data point a2, Norm() is a normalization function, is a coordinate Euclidean distance between data point a1 and data point a2, is a gray value corresponding to data point a1, is a gray value corresponding to data point a2; and is larger, is larger, and vice versa and is smaller, is smaller, is smaller, which means that the gray of two data points is more similar and the spatial position distance is closer, and the two data points are more likely to be clustered into the same cluster.
[0025] After obtaining the metric distance between data points in the three-dimensional space, all data points in the three-dimensional space are clustered by K-means clustering based on the metric distance between data points in the three-dimensional space to obtain each cluster. The clustering process using the K-means clustering algorithm under the premise of knowing the metric distance between data points is a known technology, and therefore this embodiment will not be described. The clustering result obtained based on the metric distance considers both the distance distribution of data points in the space and the similarity of the gray values of data points, and data points with similar spatial positions and gray values can be clustered into a cluster, i.e., data points in the same cluster have similar spatial positions and gray values.
[0026] Since the run-length encoding has higher compression efficiency for longer repeated similar data, this embodiment needs to make the continuous repeated or similar part in the scanning traversal path as long as possible. Therefore, after obtaining the clustering clusters, this embodiment needs to determine the primary, secondary, and final directions of scanning traversal based on the distribution characteristics of clustering, and then perform scanning traversal based on the obtained primary, secondary, and final directions of scanning traversal. The scanning traversal based on the primary, secondary, and final directions of scanning traversal can make the gray similar points in the same frame and the gray similar points between different frames continuous in the sequence obtained by scanning traversal as much as possible, so as to improve the coding compression effect.
[0027] Since the longer the comprehensive length of the same type of edge length of the circumscribed cube of all clustering clusters in the space, the more the corresponding type of edge direction can reflect the distribution trend of the overall data or gray scale in the space, and the number of continuous repeated or similar points in the final sequence obtained will be relatively more when taking this direction as the main scanning direction relative to other directions as the main scanning direction, so the embodiment will determine the primary and secondary final direction of scanning and traversal according to the edge length of the circumscribed cube of the clustering cluster; but the distribution of data points in the clustering cluster in the circumscribed cube of the clustering cluster is different, and there may be some clustering clusters with relatively discrete data point distribution but large overall distribution range, and the degree of continuous repetition or continuous similarity of data in such clustering clusters is poor, and if the primary and secondary final direction of scanning and traversal is determined, too much reference or too much belief in the edge length of the circumscribed cube of such clustering clusters will lead to low reliability of the determined primary and secondary final direction, resulting in low degree of continuous repetition or continuous similarity of data in the actual scanning and traversal path, so the embodiment also considers the actual distribution of data points in the clustering cluster when determining the primary and secondary final direction of scanning and traversal according to the edge length of the circumscribed cube of the clustering cluster, that is, the embodiment needs to obtain the direction confidence of each clustering cluster according to the number of neighborhood data points of data points in each clustering cluster and the proportion of cluster data points in the minimum circumscribed cube of each clustering cluster before determining the primary and secondary final direction of scanning and traversal according to the edge length of the circumscribed cube of the clustering cluster. The direction confidence of each clustering cluster can reflect the reference degree or importance of the corresponding clustering cluster when determining the primary and secondary final direction of scanning and traversal, and the specific process of obtaining the direction confidence of each clustering cluster is as follows:
[0028] Firstly, the minimum circumscribed cube of each cluster is constructed in three-dimensional space, and the embodiment requires that the edges of the minimum circumscribed cube of each cluster constructed need to be parallel to the coordinate axes of the three-dimensional space, that is, the length of the minimum circumscribed cube of each cluster is parallel to the horizontal axis of the three-dimensional space, the width of the minimum circumscribed cube is parallel to the vertical axis of the three-dimensional space, and the height of the minimum circumscribed cube is parallel to the vertical axis of the three-dimensional space; then the total number of data points in each cluster, the volume of the minimum circumscribed cube of each cluster, the total number of data points belonging to the preset neighborhood of the data point and belonging to the same cluster as the corresponding data point are obtained, the calculation process of the volume of the cube is known, and the filling degree representation value of each cluster is obtained according to the ratio of the total number of data points in each cluster to the volume of the minimum circumscribed cube of the corresponding cluster, that is, the filling degree representation value of any cluster is the ratio of the total number of data points in the cluster to the volume of the minimum circumscribed cube of the cluster, then the intra-cluster distribution density representation value of each cluster is obtained according to the total number of data points belonging to the preset neighborhood of the data point and belonging to the same cluster as the corresponding data point, and finally the sum of the filling degree representation value of each cluster and the intra-cluster distribution density representation value of the corresponding cluster is recorded as the directional confidence of the corresponding cluster.
[0029] The specific acquisition process of the intra-cluster distribution density representation value of each cluster according to the total number of data points belonging to the preset neighborhood of the data point and belonging to the same cluster as the corresponding data point is as follows:
[0030] The maximum total number of data points in the preset neighborhood of the data point is obtained and recorded as the total neighborhood number value, the maximum total number of data points in the preset neighborhood of each data point is the same, and in specific applications, the implementer needs to set the preset neighborhood according to the actual situation, such as setting the preset neighborhood to 26 neighborhoods in this embodiment, so the maximum total number of data points in the preset neighborhood of any data point is 26, and the total neighborhood number value is also 26; for any data point, the total number of data points belonging to the preset neighborhood of the data point and belonging to the same cluster as the data point is obtained and recorded as the intra-cluster neighborhood number value of the data point, that is, if there are 5 data points in the cluster to which the data point belongs that belong to the 26 neighborhoods of the data point, then the intra-cluster neighborhood number value of the data point is 5, the ratio of the intra-cluster neighborhood number value of the data point to the total neighborhood number value is calculated and recorded as the intra-cluster neighborhood proportion of the data point; the mean of the intra-cluster neighborhood proportion of all data points in each cluster is calculated and recorded as the intra-cluster distribution density representation value of the corresponding cluster.
[0031] And the specific calculation expression of the directional confidence of any cluster q is:
[0032]
[0033] Wherein, is the directional confidence of the cluster q, is the total number of data points in the clustering cluster q, is the volume of the minimum circumscribed cube of the clustering cluster q, is the filling degree representation value of the clustering cluster q, is the intra-cluster neighborhood number value of the i-th data point in the clustering cluster q, and N0is the total neighborhood number value, is the intra-cluster distribution density representation value of the clustering cluster q. Since when the data points in the clustering cluster occupy a larger proportion relative to the minimum circumscribed cube of the corresponding clustering cluster, that is, the filling degree of the data points in the corresponding clustering cluster in the minimum circumscribed cube of the clustering cluster is higher, it indicates that the minimum circumscribed cube of the corresponding clustering cluster can better represent the corresponding clustering cluster itself, and also indicates that the corresponding clustering cluster is more important in subsequent determination of the primary, secondary and terminal directions of the scan traversal, so when is larger, it indicates that the minimum circumscribed cube of the clustering cluster q can better represent the clustering cluster q itself, and also indicates that the clustering cluster q is more important in subsequent determination of the primary, secondary and terminal directions of the scan traversal, and the more important the corresponding clustering cluster is, the higher the direction confidence corresponding to the corresponding clustering cluster is; when the vacancies in the clustering cluster are more or the data points in the clustering cluster are more dispersed, the continuity and gray level repetition similarity of the data points in the corresponding clustering cluster will be weaker, and when the data points in the clustering cluster are more closely distributed and the number of vacancies between the data points in the clustering cluster is smaller, the continuity and gray level repetition similarity of the data points in the corresponding clustering cluster will be stronger, indicating that the corresponding clustering cluster is more important in subsequent determination of the primary, secondary and terminal directions of the scan traversal, so when is larger, it indicates that the data points in the clustering cluster q are more closely distributed and the number of vacancies between the data points in the clustering cluster q is smaller, and also indicates that the clustering cluster q is more important in subsequent determination of the primary, secondary and terminal directions of the scan traversal, and the more important the corresponding clustering cluster is, the higher the direction confidence corresponding to the corresponding clustering cluster is; and when and are larger, are larger, so when is larger, the clustering cluster q is more important in subsequent determination of the primary, secondary and terminal directions of the scan traversal, and the more important the clustering cluster q is in subsequent determination of the primary, secondary and terminal directions of the scan traversal, the more the clustering cluster q needs to be referred to in subsequent determination of the primary, secondary and terminal directions of the scan traversal or the more credible the clustering cluster q is in subsequent determination of the primary, secondary and terminal directions of the scan traversal.
[0034] After obtaining the direction confidence corresponding to each clustering cluster, the embodiment obtains the primary direction of scan traversal, the secondary direction of scan traversal and the terminal direction of scan traversal according to the direction confidence corresponding to each clustering cluster and the edge length of the minimum circumscribed cube of each clustering cluster, and the primary direction of scan traversal, the secondary direction of scan traversal and the terminal direction of scan traversal are the basis for subsequent scan traversal, and the specific determination process of the primary direction of scan traversal, the secondary direction of scan traversal and the terminal direction of scan traversal is as follows:
[0035] First, according to the direction confidence corresponding to each cluster and the edge length of the minimum circumscribed cube of each cluster, the weighted length, the weighted width and the weighted height of the minimum circumscribed cube of each cluster are obtained, the weighted length of the minimum circumscribed cube of any cluster is the result of multiplying the length of the minimum circumscribed cube of the cluster by the direction confidence corresponding to the cluster, the weighted width of the minimum circumscribed cube of any cluster is the result of multiplying the width of the minimum circumscribed cube of the cluster by the direction confidence corresponding to the cluster, and the weighted height of the minimum circumscribed cube of any cluster is the result of multiplying the height of the minimum circumscribed cube of the cluster by the direction confidence corresponding to the cluster; then the weighted length of the minimum circumscribed cube of all clusters is accumulated, and the accumulation result is recorded as the representative value corresponding to the horizontal axis direction, the long side of the minimum circumscribed cube is parallel to the horizontal axis of the three-dimensional space, the weighted width of the minimum circumscribed cube of all clusters is accumulated, and the accumulation result is recorded as the representative value corresponding to the vertical axis direction, the wide side of the minimum circumscribed cube is parallel to the vertical axis of the three-dimensional space, and the weighted height of the minimum circumscribed cube of all clusters is accumulated, and the accumulation result is recorded as the representative value corresponding to the vertical axis direction, the high side of the minimum circumscribed cube is parallel to the vertical axis of the three-dimensional space; and the specific expression of the representative value corresponding to the horizontal axis direction is:
[0036]
[0037] wherein L is the representative value corresponding to the horizontal axis direction, V0 is the total number of clusters, is the length of the minimum circumscribed cube of the jth cluster, is the direction confidence corresponding to the jth cluster, is greater, the greater the participation degree of the length of the minimum circumscribed cube of the jth cluster when calculating L, and L is positively correlated, and the expressions of the other representative values are the same.
[0038] Then, the horizontal axis direction, the vertical axis direction and the vertical axis direction are sorted in descending order of the representative values, the sorting result is recorded as the direction sequence, the first direction in the direction sequence is selected as the main scanning traversal direction, the second direction in the direction sequence is selected as the secondary scanning traversal direction, and the three directions in the direction sequence are selected as the end scanning traversal direction, for example, if the direction sequence is {vertical axis direction, horizontal axis direction, vertical axis direction}, then the vertical axis direction is the main scanning traversal direction, the horizontal axis direction is the secondary scanning traversal direction, and the vertical axis direction is the end scanning traversal direction, and the direction with a larger representative value is closer to the distribution trend of the overall data or gray scale in the space, and when subsequent scanning traversal is performed, the direction with the largest representative value is the main scanning traversal direction, and compared with other directions as the main scanning traversal direction, the number of continuous repeated or similar points in the final obtained sequence will be relatively more.
[0039] The compression storage management module 03 is configured to encode the data points in the three-dimensional space according to the main scanning direction, the secondary scanning direction and the terminal scanning direction to obtain compressed data, and to store and manage the compressed data.
[0040] After the main scanning direction, the secondary scanning direction and the terminal scanning direction are obtained, the data points in the three-dimensional space are encoded according to the main scanning direction, the secondary scanning direction and the terminal scanning direction to obtain compressed data, and the specific process of encoding the data points in the three-dimensional space according to the main scanning direction, the secondary scanning direction and the terminal scanning direction to obtain compressed data is as follows:
[0041] The data points in the three-dimensional space are scanned according to the main scanning direction, the secondary scanning direction and the terminal scanning direction to obtain a scanning sequence, and the specific process of scanning the data points in the three-dimensional space according to the main scanning direction, the secondary scanning direction and the terminal scanning direction to obtain a scanning sequence is as follows: the center point of the coordinate axis of the three-dimensional space is taken as the starting point of scanning, and the center point is the origin in the three-dimensional space; the data points in the three-dimensional space are scanned along the main scanning direction from the starting point of scanning; when scanning reaches the end, the scanning position is translated by one unit length in the secondary scanning direction, and the scanning of the data points in the three-dimensional space along the main scanning direction is continued; the above process of scanning along the main scanning direction and translating in the secondary scanning direction is repeated until the secondary scanning direction is translated to the end and the main scanning direction is scanned to the end; the scanning position is translated by one unit length in the terminal scanning direction, and the scanning of the data points in the three-dimensional space along the main scanning direction is continued; the above entire process is repeatedly continued until all the data points in the three-dimensional space are scanned, and the scanning sequence is arranged according to the scanning order; the sequence obtained by the arrangement is recorded as the scanning sequence.
[0042] If there are three frames of images in the video to be compressed, the number of rows and columns in each frame of image is three, the main direction of scan traversal is the horizontal axis direction, the secondary direction of scan traversal is the vertical axis direction, the end direction of scan traversal is the longitudinal axis direction, and the data point corresponding to the origin in the three-dimensional space is mapped from the lower left corner pixel point of the first frame of image in the video to be compressed, then the scan traversal of the first row of the first frame of image, the first row of the second frame of image, and the first row of the third frame of image will be completed in sequence from the origin, then the scan traversal of the second row of the first frame of image, the second row of the second frame of image, and the second row of the third frame of image will be completed in sequence, and finally the scan traversal of the third row of the first frame of image, the third row of the second frame of image, and the third row of the third frame of image will be completed in sequence. The data points in the three-dimensional space are arranged according to the order of scan traversal to obtain a scan sequence.
[0043] After obtaining the scan sequence, the sequence of the difference gray values to be obtained is obtained according to the scan sequence, the sequence of the difference gray values to be obtained is a sequence composed of the gray values corresponding to the data points in the scan sequence, that is, the gth gray value in the sequence of the difference gray values to be obtained is the gray value corresponding to the gth data point in the scan sequence, then the sequence of the difference gray values to be obtained is processed by difference, the difference sequence of the sequence of the difference gray values to be obtained is obtained, the result of subtracting the (h+1)th data from the hth data in the sequence of the difference gray values to be obtained is the hth data in the difference sequence; then the difference sequence of the sequence of the difference gray values to be obtained is run-length encoded, and the obtained encoding data is recorded as compressed data, and the main, secondary and end directions of scan traversal determined based on the above process can make the degree of continuous repetition or continuous similarity of the obtained scan sequence as high as possible, because there are still more similar but not identical continuous data in the sequence of the difference gray values to be obtained, and the above difference processing can convert more gray similar data into gray identical data, which can further improve the compression effect.
[0044] After obtaining the compressed data of the video to be compressed, the compressed data of the video to be compressed can be stored and managed.
[0045] Thus, the compression and storage management of the video to be compressed is completed, and the compression effect is improved, so as to save storage space and reduce storage cost, that is, to achieve effective compression and storage management of video data.
[0046] To sum up, the embodiment comprises a video mapping module, which is configured to obtain a to-be-compressed video, and map each frame image in the to-be-compressed video to a data point in a three-dimensional space; a scanning traversal direction obtaining module, which is configured to obtain a metric distance between data points in the three-dimensional space according to a spatial distance and a gray difference between the data points, and cluster the data points in the three-dimensional space according to the metric distance between the data points to obtain each cluster, and obtain a main scanning traversal direction, a secondary scanning traversal direction and a terminal scanning traversal direction according to a number of neighborhood data points of the data points in each cluster, a side length of a minimum circumscribed cube of each cluster and a proportion of intra-cluster data points in the corresponding minimum circumscribed cube; and a compression storage management module, which is configured to scan and traverse the data points in the three-dimensional space according to the main scanning traversal direction, the secondary scanning traversal direction and the terminal scanning traversal direction to obtain compressed data, and store and manage the compressed data. The main scanning traversal direction, the secondary scanning traversal direction and the terminal scanning traversal direction determined according to the cluster distribution characteristics can make the sequence obtained by scanning and traversal have a higher degree of continuous repetition or continuous similarity, thereby improving the compression storage management effect of the video.
[0047] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not limit the same; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and 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 management platform with audio and video local storage management function, characterized in that, The management platform with the audio and video local storage management function comprises: a video mapping module, configured to obtain a to-be-compressed video, and map each frame image in the to-be-compressed video into a three-dimensional space to obtain data points in the three-dimensional space; a scanning traversal direction obtaining module, configured to obtain a metric distance between data points in the three-dimensional space according to a spatial distance and a gray difference between the data points, and cluster the data points in the three-dimensional space according to the metric distance to obtain each cluster, and obtain a scanning traversal main direction, a scanning traversal secondary direction and a scanning traversal end direction according to a number of neighborhood data points of the data points in each cluster, a side length of a minimum circumscribed cube of each cluster and a cluster-in-data-point proportion in a corresponding minimum circumscribed cube; a compression storage management module, configured to scan and traverse the data points in the three-dimensional space according to the scanning traversal main direction, the scanning traversal secondary direction and the scanning traversal end direction to obtain compressed data, and store and manage the compressed data; the method for obtaining the scanning traversal main direction, the scanning traversal secondary direction and the scanning traversal end direction comprises: obtaining a direction confidence degree corresponding to each cluster according to a number of neighborhood data points of the data points in each cluster and a cluster-in-data-point proportion in a minimum circumscribed cube of each cluster, the length of the minimum circumscribed cube being parallel to a horizontal axis of the three-dimensional space, the width of the minimum circumscribed cube being parallel to a vertical axis of the three-dimensional space, and the height of the minimum circumscribed cube being parallel to a vertical axis of the three-dimensional space; and obtaining the scanning traversal main direction, the scanning traversal secondary direction and the scanning traversal end direction according to the direction confidence degree corresponding to each cluster and a side length of the minimum circumscribed cube of each cluster; the method for obtaining the direction confidence degree corresponding to each cluster comprises: obtaining a filling degree characteristic value and a cluster-in-distribution density characteristic value of each cluster according to a ratio of a total number of data points in each cluster to a volume of a minimum circumscribed cube of the corresponding cluster and a total number of data points belonging to a preset neighborhood of the data points and belonging to the same cluster as the corresponding data points, and taking a sum of the filling degree characteristic value of each cluster and a cluster-in-distribution density characteristic value of the corresponding cluster as the direction confidence degree corresponding to the corresponding cluster; the ratio of the total number of data points in each cluster to the volume of the minimum circumscribed cube of the corresponding cluster is the filling degree characteristic value of the corresponding cluster; the method for obtaining the cluster-in-distribution density characteristic value comprises: taking a total number of data points in a preset neighborhood of the data points as a total neighborhood number value, taking a total number of data points belonging to a preset neighborhood of the data points and belonging to the same cluster as the corresponding data points as a cluster-in-neighborhood number value of the corresponding data points, taking a ratio of the cluster-in-neighborhood number value of the data points to the total neighborhood number value as a cluster-in-neighborhood proportion of the corresponding data points, and taking an average of cluster-in-neighborhood proportions of all data points in the cluster as the cluster-in-distribution density characteristic value of the corresponding cluster. According to the direction confidence corresponding to each cluster and the edge length of the minimum circumscribed cube of each cluster, a method for obtaining a main scanning direction, a secondary scanning direction and a tail scanning direction comprises: The cumulative result of the weighted length of the minimum circumscribed cube of each cluster is recorded as a representative value corresponding to the horizontal axis direction, the cumulative result of the weighted width of the minimum circumscribed cube of each cluster is recorded as a representative value corresponding to the vertical axis direction, and the cumulative result of the weighted height of the minimum circumscribed cube of each cluster is recorded as a representative value corresponding to the vertical axis direction. The weighted length of the minimum circumscribed cube of any cluster is the product of the length of the minimum circumscribed cube of the corresponding cluster and the direction confidence corresponding to the corresponding cluster. The weighted width of the minimum circumscribed cube of any cluster is the product of the width of the minimum circumscribed cube of the corresponding cluster and the direction confidence corresponding to the corresponding cluster. The weighted height of the minimum circumscribed cube of any cluster is the product of the height of the minimum circumscribed cube of the corresponding cluster and the direction confidence corresponding to the corresponding cluster. The horizontal axis direction, the vertical axis direction and the vertical axis direction are sorted in descending order of the representative values, and a direction sequence is obtained. The first direction in the direction sequence is taken as the main scanning direction, the second direction in the direction sequence is taken as the secondary scanning direction, and the three directions in the direction sequence are taken as the tail scanning direction.
2. The management platform with audio and video local storage management function according to claim 1, characterized in that, The method for obtaining the metric distance between data points in the three-dimensional space comprises: For data point a1 and data point a2 in the three-dimensional space, the normalized result of the coordinate Euclidean distance between the data point a1 and the data point a2 is recorded as a first distance value, the normalized result of the absolute value of the gray difference value between the data point a1 and the data point a2 is recorded as a second distance value, and the sum of the first distance value and the second distance value is taken as the metric distance between the data point a1 and the data point a2.
3. The management platform with audio and video local storage management function according to claim 1, characterized in that, The method for obtaining compressed data by scanning and traversing the data points in the three-dimensional space according to the main scanning direction, the secondary scanning direction and the tail scanning direction comprises: According to the main scanning direction, the secondary scanning direction and the tail scanning direction, the data points in the three-dimensional space are scanned and traversed to obtain a scanning sequence. The sequence of gray values corresponding to each data point in the scanning sequence is recorded as a sequence of to-be-difference gray values. The difference sequence of the to-be-difference gray value sequence is run-length encoded to obtain compressed data.
4. The management platform with audio and video local storage management function according to claim 3, characterized in that, The method for obtaining a scanning sequence by scanning and traversing the data points in the three-dimensional space according to the main scanning direction, the secondary scanning direction and the tail scanning direction comprises: Taking the origin in the three-dimensional space as a scanning traversal starting point, data points in the three-dimensional space are sequentially scanned along a main scanning traversal direction from the scanning traversal starting point, when scanning traversal is performed to an end, the scanning position is translated by a unit length in a secondary scanning traversal direction, and scanning traversal of the data points in the three-dimensional space along the main scanning traversal direction is continued, the process of scanning traversal along the main scanning traversal direction and translation in the secondary scanning traversal direction is repeatedly performed until the translation in the secondary scanning traversal direction is performed to an end, then the scanning traversal position is translated by a unit length in a final scanning traversal direction, and scanning traversal of the data points in the three-dimensional space along the main scanning traversal direction is continued, the whole process is repeatedly performed until all the data points in the three-dimensional space are scanned, and the scanning is ended, and a sequence obtained by arranging the data points in the three-dimensional space according to the scanning sequence is recorded as a scanning sequence.
5. The management platform with audio and video local storage management function according to claim 1, characterized in that, The horizontal axis of the three-dimensional space represents a horizontal pixel position in a single frame image, the vertical axis represents a vertical pixel position in the single frame image, and the vertical axis represents a sequence number of the image in a video to be compressed.
Citation Information
Patent Citations
Data compression storage method of computer system
CN115834887A
Edge computing-oriented AI gateway video efficient compression method
CN120676183A