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

Multi-target tracking method based on structured light and SiamMask network

A multi-target tracking and structured light technology, applied in the tracking field, can solve problems such as poor precision, inconvenient application, and inability to automatically obtain target coordinates for single target tracking, and achieve the effects of convenient use, improved accuracy and practicability

Active Publication Date: 2019-12-06
YANSHAN UNIV
View PDF6 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003]Single target tracking cannot automatically obtain the target coordinates, and can only identify one target, which is not convenient to apply to real life; the multi-target tracking network uses the algorithm of correlation filtering as the tracker, The accuracy is relatively poor; neither uses a camera with structured light

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 structured light and SiamMask network
  • Multi-target tracking method based on structured light and SiamMask network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0026] Embodiment 1: a kind of multi-target tracking method based on structured light and SiamMask network, comprises the following steps:

[0027] A. Use the realsense D435 depth camera to obtain the color image and the corresponding depth image, and call the pyrealsense2 library function to align the color image with the depth image;

[0028] B. The coordinates of the target frame are generated, using the target detection model, the improved network based on yolov3, receiving the color image and the depth image, obtaining their features respectively and performing feature fusion after the final convolution layer, and combining the fused features Figure 1 It is sent to the fully connected layer for classification and regression, and the frame coordinates of the generated target are sent to the SiamMask network:

[0029] C. The SiamMask network receives the target coordinates and starts tracking. Multiple targets generate multiple feature maps, each feature map is compared wi...

Embodiment 2

[0032] Embodiment 2, on the basis of Embodiment 1, step E is specifically to select different targets with borders of different colors, and add a specific ID or name to each target, so as to distinguish each target.

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 structured light and a SiamMask network. The multi-target tracking method includes the steps: introducing structured light to obtain a depth image to further improve the detection precision while entering a target detection network to automatically extract target position coordinates, improving a single-target tracking network SiamMask to realize multi-target tracking, providing a matching algorithm between detection and tracking and adopting a deep neural network Resnet for re-identification, so as to guarantee that multiple targets are detected and tracked at the same time. According to the multi-target tracking method, a correlation filtering algorithm in a traditional multi-target tracking network is replaced, and a SiamMask network based on a deep neural network is adopted. The SiamMask network is the network with the highest precision at present in video tracking, when the IOU threshold value is 0.5, the precision ona VOT data set reaches 90%, and meanwhile, depth image information is added into the algorithm, so that the detection precision is further improved. In order to improve the network operation speed, the algorithm abandons a mask branch of a SiamMask network, and the mask branch can be directly added in the later period to present a segmented image.

Description

technical field [0001] The invention relates to a tracking method, in particular to a multi-target tracking method based on structured light and a SiamMask network. Background technique [0002] Existing tracking networks include single-target tracking networks such as SiamMask networks and multi-target tracking networks such as deep_sort_yolov3 networks. [0003] Single-target tracking cannot automatically obtain target coordinates, and can only identify one target, which is not convenient for application in real life; multi-target tracking networks use correlation filtering algorithms as trackers, and the accuracy is relatively poor; neither of them use structured light cameras. Contents of the invention [0004] The purpose of the present invention is to provide a multi-target tracking method based on structured light and SiamMask network to solve the problems raised in the background technology above. [0005] To achieve the above object, the present invention provide...

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
IPC IPC(8): G06T7/246G06T7/33G06T7/38G06N3/04G06N3/08
CPCG06T7/246G06T7/33G06T7/38G06N3/08G06T2207/10024G06T2207/10028G06T2207/20081G06T2207/20084G06N3/045
Inventor 呼子宇高博马学敏宋浩诚
Owner YANSHAN UNIV
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