Video grading code stream coding method and system based on neural network processor array

Abstract

The invention provides a video hierarchical code stream coding method based on a neural network processor array. The method comprises the following steps: obtaining synchronous videos, and creating video objects in the synchronous videos according to a preset analysis model; obtaining the synchronous videos, and optimizing images in the synchronous videos according to a preset optimization model so as to create optimized videos; obtaining the video objects, task processing models and the synchronous videos, allocating processing resources to parallel task processing procedures of the corresponding task processing models, and carrying out corresponding task processing for the video objects in a parallel manner, so as to create corresponding object description information; and carrying out video coding of the optimized videos, carrying out coding of the corresponding object description information, and embedding the coded object description information into video code streams, so as to create hierarchical code streams of a source video. The method provided by the invention has the advantages that analysis and processing result information of the optimized videos and the various video objects are synchronously combined through coding, so that the hierarchical code streams containing the relevant information of video contents and the various video objects are created for the source video. The invention further provides a video hierarchical code stream coding architecture based on the neural network processor array.

Description

technical field [0001] The present invention relates to a data processing method and processing device, in particular to a display data processing method and processing device. Background technique [0002] Video transmission consumes a lot of bandwidth. In the field of video surveillance, real-time high-definition streams consume the most bandwidth and system processing resources. Therefore, in the prior art, the bandwidth problem is usually solved by encoding and compressing the source video to form a low-rate bit stream suitable for real-time transmission and a high-rate bit stream suitable for delayed transmission. However, only using the code rate to adapt the bandwidth cannot guarantee the real-time and reliable transmission of information in the video. When the code rate is reduced, the quality of information carried in the video is correspondingly reduced or even unusable. It is feasible to use different encoding methods to encode the information in the video and t...

AI Technical Summary

Problems solved by technology

[0005] In view of this, the embodiment of the present invention provides a video hierarchical code stream encoding method based on a neural network processor array, which solves the technical problem in the existing video encoding technology that the encoding rate cannot be overcome and the quality of the video information is reduced.

[0006] The embodiment of the present invention also provides a video hierarchical code stream encoding system based on a neural network processor array, which solves the technical problem that the existing video encoding architecture is unreasonable, resulting in the inability to perform real-time load balancing on the information carried in the video. Enables parallel processing of multiple detection and recognition tasks

Images