Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Multi-target tracking method based on multi-agent deep reinforcement learning

A multi-target tracking and reinforcement learning technology, applied in the field of multi-target tracking based on multi-agent deep reinforcement learning, can solve problems such as frequent occlusion, and achieve the effect of ensuring accuracy, fewer false negatives, and guaranteed speed

Active Publication Date: 2021-09-07
NANJING QIANHE INTERNET OF THINGS TECH CO LTD
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Due to a large number of influencing factors in video scenes, such as: the appearance and disappearance of objects, frequent occlusion between objects, similar appearance of multiple objects, and background noise, etc., multi-object tracking is still a challenging research topic.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Multi-target tracking method based on multi-agent deep reinforcement learning
  • Multi-target tracking method based on multi-agent deep reinforcement learning
  • Multi-target tracking method based on multi-agent deep reinforcement learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0018] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0019] A multi-target tracking method based on multi-agent deep reinforcement learning, comprising the following steps:

[0020] (1) if figure 1 As shown, the YOLO V3 target detector is used to detect multiple targets in each frame of the video to be tested; for the tth frame of image, the output result of the target detector is the set D t , set D t Contains the detection results of multiple targets, the detection results are displayed by the target box, and the detection results of a single target are recorded as d t =(x, y, w, h), (x, y) is the coordinates of the center point of the target frame, w, h are the width and height of the target frame respectively;

[0021] (2) Define the following parameters: each detected target is regarded as an agent, expressed as agent i, i∈I≡{1,...,n}, n is the number of agents; Each frame of image is re...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a multi-target tracking method based on multi-agent deep reinforcement learning. Multiple targets are detected by a target detector, and the detected multi-targets are regarded as multiple intelligent agents, and then the deep reinforcement learning method is used to obtain multiple targets. The joint action set of the target, and then complete the multi-target tracking. The present invention applies the multi-agent deep reinforcement learning technology to the multi-target tracking method for the first time, which can overcome the technical shortcomings that the artificially designed features are not comprehensive enough and not accurate enough, and at the same time can increase the calculation speed, realize real-time tracking, and have high multi-target tracking Accuracy, precision, less false positives and false positives, less affected by various interference factors in multi-target tracking scenarios, and more accurate tracking results.

Description

technical field [0001] The invention relates to a video target tracking method, in particular to a multi-target tracking method based on multi-agent deep reinforcement learning. Background technique [0002] As a hot issue in the field of computer vision, multi-target tracking based on video has been widely used in many application fields, such as: automatic driving, robot navigation, artificial intelligence, etc. Due to a large number of influencing factors in video scenes, such as: the appearance and disappearance of targets, frequent occlusions between targets, similar appearance of multiple targets, and background noise, etc., multiple target tracking is still a challenging research topic. . Contents of the invention [0003] Purpose of the invention: In order to overcome the influence of a large number of interference factors in the prior art on multi-target tracking, the present invention provides a multi-target tracking method based on multi-agent deep reinforcemen...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06T7/20G06N3/04G06N3/08
CPCG06T7/20G06T2207/10016G06T2207/20081G06T2207/20084
Inventor 姜明新季仁东荣康王国达陈寒章
Owner NANJING QIANHE INTERNET OF THINGS TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com