Night target detecting and tracing method based on visual property

Abstract

The invention relates to a method for detecting and tracking a target of the night, which includes following steps: it captures video of the night, and then it determines the adjacent frames difference based on adaptive algorithm, and then it determines the target region according to the contrast, and then it forecasts the current location and the next time location according to the speed of target. The invention can accurately judges the similarity frames, which is helpful for the follow-up process of treatment. The algorithm makes the video not only in the evening but also in bad weather conditions such as fog can be effective.

Description

technical field [0001] The invention relates to pattern recognition, in particular to a vision-based night target detection and tracking method. Background technique [0002] With the rapid development of modern science and technology, the use of cameras to monitor dynamic scenes has been widely used in all aspects of modern society, especially those occasions that are sensitive to security requirements, such as national defense, communities, banks, parking lots, military bases, etc. Visual monitoring of dynamic scenes is a frontier research direction that has attracted much attention in recent years. It detects, recognizes, tracks targets and understands their behavior from image sequences captured by cameras. Although the surveillance camera, which is an extension of human vision, has been widely used in commercial applications, it has not fully utilized its role as a real-time and active surveillance medium. Therefore, it is increasingly urgent and necessary to develop a...

AI Technical Summary

Problems solved by technology

But the high price still prevents it from being widely used

From the perspective of visible light, there is not much research on the automatic detection of targets in videos captured by ordinary CCD cameras.

Images