Font Size: a A A

The Research And Implementation Of Surveillance Video Compression Algorithm Based On Redundant Dictionary

Posted on:2016-02-09Degree:MasterType:Thesis
Country:ChinaCandidate:Y H BaiFull Text:PDF
GTID:2308330503950750Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Surveillance video system as an effective means of public safety widely appeared in people’s life in recent years. It produces huge amounts of surveillance video data, to storage and transport has brought great pressure. Traditional video coding technology based on combining orthogonal transformation and motion estimation of hybrid coding structure, the encoding process is complex, computation, and conflicts with hardware limitations of the monitoring system. Surveillance video is the background for a long time is relatively fixed, outlook change often has the characteristics of the video. Common video coding algorithm, rarely for monitoring video special custom, it is difficult to more targeted surveillance video compression, so the study of surveillance video coding technology need breakthrough.This article is based on signal sparse representation theory of redundant dictionary surveillance video compression coding process, was studied based on the redundant dictionary surveillance video compression algorithm. In view of the traditional hybrid coding structure compensation of the high complexity of motion estimation, explore compression algorithm of motion estimation has been distributed, specific include:Research and implements the monitoring video sparse decomposition algorithm based on redundant dictionary, on the basis of analyzing the characteristics of monitoring video, a redundant dictionary based on image block design and training method, and on this basis, designs and realizes a kind of surveillance video sparse decomposition algorithm based on redundant dictionary. Experimental results show that the design method of the redundant dictionary designed for each image block to establish effective redundant dictionary, based on the established dictionary can achieve effective description of surveillance video, and the reconstruction effect is good.Research and implements the monitoring video compression coding algorithm based on redundant dictionary. Based on the established the redundant dictionary design method based on image block, designs and realizes a redundant dictionary based and JPEG-LS, on the basis of the sparse decomposition of surveillance video, the sparse decomposition coefficient and redundant dictionary respectively founded by quantization and compression coding, implements the effective of surveillance video compression. For further training of the proceeds of the dictionary to prospect target sparse representation ability is weak, this paper proposes a compression algorithm based on key frames and key frames keep outlook, sparse decomposition, quantification and coding to the background, and non-critical frame retain only prospect. Experimental results show that the algorithm is based on key frames in high compression ratio can significantly improve the quality of reconstruction.Research and design implements a dictionary of surveillance video compression algorithm based on area. According to the method based on image block computation, dictionary training time problem, put forward a kind of based on image region redundant dictionary design and training method, on top of this dictionary is designed and implemented based on the area of monitoring video compression algorithm. Dictionary of surveillance video sparse decomposition algorithm using area, the rest of the processing reference monitor video compression coding algorithm based on redundant dictionary. The experimental results show that the monitoring of key frames based on regional dictionary and video compression algorithm to the reconstruction quality under the condition of equal, compression ratio than MJPEG increased by 1.5 to 2.3 times.
Keywords/Search Tags:Surveillance video, Sparse decomposition, Redundant dictionary, Video coding
PDF Full Text Request
Related items