A Single Camera Pedestrian Tracking Method

A pedestrian tracking, single camera technology, applied in the field of single camera pedestrian tracking, can solve the problem of interruption of pedestrian trajectories and achieve the effect of improving target confusion

Active Publication Date: 2019-12-03
SUN YAT SEN UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This strategy can well solve the problem of trajectory interruption caused by short-term occlusion of pedestrians during the tracking process.

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 Pedestrian Tracking Method
  • A Single Camera Pedestrian Tracking Method
  • A Single Camera Pedestrian Tracking Method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0068] The pedestrian tracking method of the single camera provided by the present invention, such as figure 1 , 3 , 4, including the following steps:

[0069] S1. The video image is detected, and the detected pedestrian is called an observation target;

[0070] S2. Assuming that there are κ observation objects detected in the current frame t, extract the features of the κ observation objects;

[0071] S3. Set in the current frame t, there are several motion trajectories before the current frame t, T={d s , d s+1 ,...,d e} represents a motion trajectory, where s and e are the starting and ending moments of the trajectory T respectively; for the motion trajectory Select the observation target that may be associated in the current frame t through the position constraints, and then calculate the motion trajectory T i The similarity with the observed target that may be associated will be related to the motion trajectory T i The observation target with the highest similarit...

Embodiment 2

[0121] In this example, the method provided in Example 1 is specifically tested. This test uses two videos of different densities in the PETS2009 database, namely S2L1 sequence and S2L2 sequence, and compares it with the current highest level method. method one [1] An energy function containing physical constraints and image evidence is constructed, and the target trajectory is found by optimizing the continuous energy function. Method Two [2] It is a method based on graph theory optimization to complete the matching of observation targets by constructing a hypergraph with undirected hierarchical associations. The performance evaluation indicators and their meanings are as follows, where "↑" indicates that the larger the value, the better the tracking result, and "↓" indicates that the smaller the value, the better the tracking result:

[0122] Label switching number (IDs↓, ID Switch), the sum of label switching times of the same real target.

[0123] The number of interrup...

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 pedestrian tracking method with a single camera. The present invention adopts a layered tracking strategy, first connects the detection targets into stable and reliable small segments of motion trajectories, and then forms the final tracking trajectory by matching the small segments of motion trajectories and filling the blank segments between the small segments of motion trajectories. This strategy is beneficial to solve the problem of trajectory interruption caused by pedestrians being occluded for a short time during the tracking process. In the construction of small segment motion trajectories, the present invention proposes a two-step matching method based on the idea of ​​cascading. Firstly, an efficient histogram matching method is used to deal with the situation where the matching difficulty is small, and then a robust sparse representation classifier is used to solve the target table Confusion caused by visual similarity. In the matching between small-segment motion trajectories, the present invention proposes a sparse representation-based trajectory distance measurement method, which has good robustness to target appearance changes caused by environmental changes and target deformations. In addition, the present invention also proposes a trajectory matching method based on hierarchical clustering with a small amount of calculation to improve the efficiency of global multi-trajectory matching.

Description

technical field [0001] The invention relates to the field of intelligent security, and more specifically, to a pedestrian tracking method with a single camera. Background technique [0002] As a promising technology in the field of security, intelligent video surveillance needs to track, detect, identify and analyze the behavior of the target of interest in the video. Multi-target tracking technology needs to detect the trajectories of multiple targets from the scene video to provide a basis for target behavior analysis. At the same time, collecting and analyzing pedestrian trajectory data from video recordings can provide more useful information for facility planning, and even discover potential commercial value. Therefore, the realization of intelligent multi-target tracking has broad application prospects. [0003] The multi-target tracking method based on multi-camera has certain advantages in solving the problem of target occlusion. Such methods generally use multi-e...

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): G06K9/00
CPCG06V40/10G06V20/41
Inventor 郑慧诚林梓健马东宏
Owner SUN YAT SEN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap