Distributed video cutting model based on big data, and application
A distributed video and big data technology, applied in special data processing applications, video data retrieval, video data indexing, etc., can solve the problems of high resource consumption, lack of data analysis foundation, low efficiency, etc., to achieve stable system operation, improve Video cutting efficiency, flexible deployment and expansion effects
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0048] The distributed video cutting model based on big data is constructed as follows:
[0049](1) Analyze the video to be cut and extract events, including passing, crossing, steals, fouls, free kicks, corner kicks, goal kicks, goals and other events, among which passing can be subdivided into passing that forms a shot , Linkage events such as the pass that forms a corner kick, the pass that forms a free kick, etc., record the event, the event start time and the event end time, and record it as physical data;
[0050] (2) According to the start and end time of the event and the start and end time of the linkage event, determine the length L of the segmented video. The formula is as follows:
[0051] L=(Ue-Ub) / 2 - (Ne-Nb) / 2
[0052] L: video cutting length
[0053] Ub: last event start time
[0054] Ue: Last event end time
[0055] Nb: next event start time
[0056] Ne: end time of the next event;
[0057] (3) Store the physical data and video in the Hadoop distributed ...
Embodiment 2
[0062] A distributed video cutting system based on big data, which consists of the following three parts:
[0063] (1) Data storage layer: adopts distributed file system architecture, composed of Hadoop, Hbase, and Oralce data storage systems, in which Hadoop is used to store video files and log files, including original game video, original physical data, game video slices, players Heat map, etc.; Hbase is used to store non-relational data, using the key-value method for data storage; Oracle is used to store business data with strong correlation;
[0064] (2) Data analysis layer: use Spark and Map-reduce to analyze the original data, analyze the start and end time of the event, occurrence coordinates, target coordinates, and player codes, and generate video cutting tasks according to the video cutting algorithm, video cutting The algorithm determines the segmented video length L according to the start and end time of the event and the start and end time of the linkage event. ...
Embodiment 3
[0074] The distributed video cutting model and the distributed video cutting system in the present invention are not independent, but a closely integrated whole, which will be described from the overall flow below.
[0075] (1) After the game or training is over, collect the game / training video and physical data through cameras, wearable devices, etc., and store the video files and physical data in the designated Hadoop storage directory;
[0076] (2) Through the event parser, analyze the original data stored in Hadoop, and extract key events including pass, cross, foul, corner kick, free kick and other event information as well as the coordinate information of the player and the ball on the field, such as the start and end time of the event , event start coordinates, target coordinates, event type, event offender, etc. After the original data analysis is completed, the converted standard data is stored in Hbase and a video cutting task is generated;
[0077] (3) The time-shar...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More