Video structured processing device and method based on FPGAs (Field Programmable Gate Array) and artificial intelligence

Abstract

The disclosure provides a video structured processing device and method based on FPGAs (Field Programmable Gate Array) and artificial intelligence. The video structured processing device (100) based on FPGAs and artificial intelligence according to the disclosure comprises a video decoder (110) for decoding the input encoded video data, and inputting the decoded video data to the FPGAs (120, 130).The FPGAs (120, 130) receive the decoded video data from the video decoder (110), perform structured processing on the received decoded video data by using an artificial intelligence algorithm programmed in the FPGA, and return the structured information to the video decoder (110). The video decoder (110) receives the structured information from the FPGAs (120, 130), processes the decoded video data according to the received structured information, and finally outputs a target picture and the corresponding structured information.

Description

technical field [0001] The present invention relates to the field of artificial intelligence, and more specifically relates to a device and method for video structured processing using FPGA. Background technique [0002] With the continuous deepening of the construction of "Safe City", "Urban Three-dimensional Public Security Prevention and Control", "Smart City" and other scientific and technological police projects, the first-tier, second-tier and third-tier cities in my country have basically completed the urban monitoring and alarm system from standard definition to high-definition. The construction of the network and the integration of video surveillance resources are being carried out based on unified national standards. First-tier and second-tier cities have developed intelligent video analysis systems on the basis of networked and integrated video resources, and have achieved preliminary results. However, as an important part of smart security and smart transportatio...

AI Technical Summary

Problems solved by technology

The main problems at present are: the lack of standardized generation methods of video information and intelligence, and the lack of a new police work mode that uses video information and intelligence to guide investigation and solve crimes; the lack of unified standards and norms for all aspects of video information and intelligence police application

[0005] (1) The analysis accuracy is greatly affected by the environment

[0006] It is difficult for the system to model complex and abnormal behaviors. The proximity of the target to the background will lead to difficulty in extracting target feature information; the occlusion of the moving target will cause the loss of target information; the target's moving speed is too fast or the algorithm is too complex, resulting in low tracking effectiveness. The above factors It is easy to cause false positives, missed negatives, and difficulty in tracking, etc.

[0007] (2) Massive data analysis speed is slow

[0008] For the structured processing of the back-end platform, the front-end acquisition equipment returns a large amount of video, pictures and other information, which needs to be superimposed on the platform server for analysis. Due to the limitation of the processing performance of the server itself, it is impossible to process a large amount of data in parallel, so , although the application of intelligent analysis server saves manpower and time for public security, traffic police and other situations that require a large amount of video retrieval, there is still a lot of room for improvement in overall efficiency

Images