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

A single-camera multi-target tracking method based on an improved graph partition model

A multi-target tracking and single-camera technology, which is applied in the field of single-camera multi-target tracking based on the improved graph partition model, can solve the problems of missed detection and wrong matching in the single-camera multi-target tracking method of the graph partition model, and reduce the number of wrong matches , Reduce the missed detection rate, improve the effect of tracking accuracy

Inactive Publication Date: 2021-04-06
HUAZHONG UNIV OF SCI & TECH
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] In view of the defects of the prior art, the purpose of the present invention is to solve the technical problems of missed detection and wrong matching in the prior art map partition model single camera multi-target tracking method

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
  • A single-camera multi-target tracking method based on an improved graph partition model
  • A single-camera multi-target tracking method based on an improved graph partition model
  • A single-camera multi-target tracking method based on an improved graph partition model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0039]In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

[0040] Single-camera multi-target tracking is performed by batch processing. Simply put, a batch of continuous pictures with several frames is input, and the appearance features and motion features are extracted from these pictures to calculate the similarity. The most similar frames of pictures are concatenated into one People's track. However, when tracking multiple targets, it is necessary to process images of an unknown number of targets at the same time, so the amount of data processed is very large.

[0041] In order to reduce the complexity of the problem to be solved, the present invention ...

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 single-camera multi-target tracking method based on an improved graph partition model, which belongs to the field of target tracking. The present invention adopts a two-layer reasoning structure, regards single-camera multi-target tracking as a graph partition problem, and uses binary integer programming (BIP) to solve it. In the stage of hierarchical reasoning, the first stage uses a shorter sliding window to divide the detection frames belonging to the same person into the same graph partition to form short trajectory segments; the second stage uses a longer sliding window to divide the detection frames belonging to the same person Small segments of individual trajectories are partitioned into the same graph partition to form long trajectories. In view of the fact that the sliding window in the second stage may miss detection near the junction of non-overlapping segments, the present invention proposes a method of overlapping sliding windows, thereby reducing the missed detection rate and improving the tracking accuracy of the algorithm. On the other hand, when multiple targets are tracked, identity switching caused by mutual occlusion between targets is prone to occur. Therefore, the present invention proposes a trajectory constraint method to reduce the occurrence of identity switching.

Description

technical field [0001] The invention belongs to the field of target tracking, and more particularly relates to a single-camera multi-target tracking method based on an improved graph partition model. Background technique [0002] Multi-target Single camera Tracking (MTSCT) is an important part of computer vision. It studies all the position information of multiple targets of interest in a continuous video frame, and obtains the motion of all targets of interest. trajectory, and then analyze the behavior of the tracking target. [0003] Today's single-camera multi-target tracking methods have achieved some good results, including Kalman filter, particle filter, correlation filter, optimal Hungarian algorithm, binary integer programming, etc. In practical application, although methods such as Kalman filter and particle filter are often intuitive and easy to understand through probability inference, they are usually difficult to infer, so the present invention often chooses op...

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): G06T7/223G06K9/62
CPCG06T7/223G06T2207/10016G06T2207/30241G06F18/231
Inventor 桑农熊月高常鑫史广亚
Owner HUAZHONG UNIV OF SCI & TECH
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