Multiple-target tracking method and device based on improved random forest

Abstract

The invention provides a multiple-target tracking method and device based on an improved random forest. The device comprises a random forest training module, a random forest classification module and a target reposition module, wherein the random forest training module is used for training and learning the target before intersection and building a target classifier; the random forest classification module is used for classifying the intersected target region through the trained classifier in the next frame; the target reposition module is used for conducting clustering operation to the classified block, so as to form the target position region. According to the method and device provided by the invention, problems of error tracking and miss tracking caused by intersection in a multiple-target tracking process, for example, the intersection of a plurality of targets in pedestrian tracking and vehicle tracking, are effectively solved.

Description

technical field [0001] The invention relates to the fields of computer vision and artificial intelligence, in particular to an improved random forest-based multi-target tracking method and device. Background technique [0002] In recent years, with the development of science and technology, the public security video surveillance system is a powerful means to actively grasp and combat urban social security, such as establishing public security video surveillance at stations, docks, airports, ports, urban traffic arteries and entrances and exits. system, giving full play to the advantages of its modern technological prevention means is of great significance to maintaining social political and public security stability. [0003] The traditional video monitoring system is only equivalent to a video capture browser. The monitoring personnel need to conduct manual observation for a long time, and only rely on subjective decision whether to take emergency measures, and the monitori...

AI Technical Summary

Problems solved by technology

[0004] An important direction in the field of intelligent image analysis is multi-target tracking. The difficulty and focus of multi-target tracking is the matching and positioning problem after multiple targets cross. Traditional methods include meanshift, optical flow, etc., and these methods cannot effectively solve target crossing. After the relocation problem, there will be a large number of matching errors, target loss and other problems. The present invention proposes a multi-target tracking system based on random forest, which can effectively reduce the probability of target tracking after target crossing. The random forest classification algorithm Because of its fast execution time, it has been applied in many fields in recent years, but the main application fields are image classificat...

Images