Supercharge Your Innovation With Domain-Expert AI Agents!

Visible light and infrared video target tracking method based on the convolution neural network

A convolutional neural network and target tracking technology, applied in biological neural network models, neural architecture, image data processing, etc., can solve problems such as motion blur and target occlusion, and achieve the effect of improving target tracking performance

Inactive Publication Date: 2018-09-11
SHANGHAI JIAO TONG UNIV
View PDF5 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The present invention aims at the above-mentioned deficiencies existing in the prior art, and proposes a visible light and infrared video target tracking method based on a convolutional neural network, a related target tracking method based on an effective convolutional neural network for visible light and infrared video fusion, using a two-layer Convolutional neural network extracts sparse features of visible light and infrared video by using convolution filters, which can effectively solve problems such as target occlusion, motion blur and complex background

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
  • Visible light and infrared video target tracking method based on the convolution neural network
  • Visible light and infrared video target tracking method based on the convolution neural network
  • Visible light and infrared video target tracking method based on the convolution neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0024] As shown in Figure 1, the essence of target tracking must learn the ability to translate changes, but the algorithm also needs to use the translation invariance of convolutional neural networks to overcome the drift problem.

[0025] In this embodiment, based on the marked first frame image, a standardized partial image is extracted from the target area of ​​the visible light and infrared image as the target convolution filter; The relevant convolution filter of the weight, so as to establish a convolutional neural network without training to perform convolution processing on the preprocessed video, and finally obtain the final tracking result after overall denoising.

[0026] In this embodiment, the correlation attributes between the local images corresponding to the image candidate windows are more reasonable for target tracking. Different from the direct tracking algorithm, the correlation tracking algorithm effectively uses the correlation between the local images c...

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 provides a visible light and infrared video target tracking method based on the convolution neural network. The method comprises the following steps that firstly, based on a marked firstframe image, a standardized local image is extracted from a target area of the visible light and the infrared image is served as a target convolution filter; the relative tracking candidate window isobtained from the visible light and the infrared image respectively to obtain the correlation convolution filter with the special weight, so that a convolution neural network which does not need to be trained is built for carrying out convolution processing on the preprocessed video; finally the final tracking result is obtained through overall denoising. According to the invention, the target tracking performance is remarkably improved.

Description

technical field [0001] The invention relates to a technology in the field of image processing, in particular to a method for tracking visible light and infrared video targets based on a convolutional neural network. Background technique [0002] Existing target tracking methods can be divided into three categories according to the selection of appearance models: generative tracking methods, discriminative tracking methods and hybrid tracking methods. The basic idea of ​​the generative tracking method is to learn an appearance model for representation tracking The target, and then the tracking method selects the candidate area most similar to the model to determine the tracking result based on the smallest reconstruction error, but the training based on the appearance model of the object does not consider the background information, so it lacks stable tracking for similar background areas ability of the target, leading to tracking drift; discriminative-based tracking methods ...

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): G06T7/246G06N3/04
CPCG06T7/246G06T2207/10048G06N3/045
Inventor 肖刚徐宁文张星辰
Owner SHANGHAI JIAO TONG UNIV
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com