Target detection method based on video image recognition technology

Abstract

The invention belongs to the technical field of image processing, and particularly relates to a target detection method based on a video image recognition technology, which comprises the following steps: video input, target monitoring, target tracking and target output. A monitoring area focused on is separated from a picture background through background modeling in area monitoring, background factor interference is reduced, invalid targets are further effectively filtered through a target area suppression method, meanwhile, effective targets are continuously and accurately positioned througha target tracking method, and false alarm targets are eliminated. Compared with the prior art, the technical scheme provided by the invention has the advantages that the background and the target inthe scene are separated by analyzing the video image of the surveillance camera without the technical support of other hardware equipment, so that the effective target appearing in the camera scene isanalyzed and tracked; the invention has the capability of filtering various backgrounds such as wind, rain, snow, fallen leaves, birds and fluttering flags, so that the practicability of the video image recognition technology is effectively improved.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a target detection method based on video image recognition technology. Background technique [0002] In recent years, with the gradual promotion and popularization of high-definition video surveillance, the video surveillance market has achieved "seeing, seeing clearly, and seeing far", and is further developing in the direction of "seeing and understanding". In order for the system to understand the monitoring content in some videos, there are mainly two traditional technology implementation methods. One is the "video linkage" technology method, which can form a linkage mechanism with other hardware devices, such as automatic correlation after triggering an alarm through an alarm device. Video, which binds the video screen with the alarm event, so that the alarm content of the corresponding target in the video surveillance can be judged. The other is the "v...

AI Technical Summary

Problems solved by technology

[0003] On the one hand, the "video linkage" technical method needs to add additional hardware equipment to realize it, which increases the construction cost of the system. At the same time, when multiple targets appear in the video, it is difficult to know which target triggers the alarm event

The "video motion detection" technology lacks a filtering mechanism for changes in the scene. When there are light changes, shadows, shaking leaves or low image quality in the picture, false alarms of target detection often occur, resulting in low accuracy of target detection. , not very practical

Images