Unlock instant, AI-driven research and patent intelligence for your innovation.

Apparent enhancement depth target tracking method

A technology of target tracking and tracking algorithm, which is applied in the field of single target tracking, can solve the problems of tracking errors and low discrimination of similar targets, achieve strong discrimination ability, balance complexity and accuracy, and enhance discrimination ability

Active Publication Date: 2021-11-02
10TH RES INST OF CETC
View PDF2 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The purpose of this invention is to address the deficiencies of the existing technology, to provide a deep target table that can reduce the complexity of development and design, and has certain versatility. Visually enhanced tracking method to solve the technical problem of low discrimination of similar targets in the prior art and cause tracking errors

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
  • Apparent enhancement depth target tracking method
  • Apparent enhancement depth target tracking method
  • Apparent enhancement depth target tracking method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0022] refer to figure 1 . According to the present invention, adopt following steps:

[0023] S1: Observe the current tracking situation of the tracking algorithm, and adopt a strategy for the change of the target scale to adaptively adjust the size of the search area; specify for the change of the target scale, the rule is to expand the search area when the target scale becomes larger, and the scale becomes smaller Narrow down the search area. When the real target becomes larger, because it will be scaled to a fixed scale, it will actually become relatively smaller, and the output features of the convolutional layer will be reduced and blurred as the scale is relatively smaller. It is necessary to continuously expand the search area to maintain sufficient The receptive field, when the real target becomes smaller, the situation is the opposite. The output features of the convolutional layer will be increased in dimension and become more specific as the scale is relatively l...

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

According to the apparent enhancement depth target tracking method disclosed by the invention, the design complexity can be reduced on the premise of not changing a backbone network structure, and the method has certain universality. The method is realized through the following technical means: searching an object similar to a template in a search area by observing the current tracking condition of a tracking algorithm, and adaptively adjusting the size of the search area according to a target scale change trend; sending the template image and the self-adaptive search area into a SiamRPN++ framework, completing the initialization of a detector, obtaining the score sorting of candidate boxes, and forming a picture library; carrying out the target matching of color distribution, morphological features and textural features on the candidate frames and the template pictures according to three kinds of traditional appearance information of colors, HOG and LBP, carrying out the similarity measurement calculation through the appearance information, obtaining the new attribute scores, carrying out the voting, acquiring the candidate frame pictures with the most similar appearance information in the second picture library, and setting the candidate frame pictures as the prediction target of the frame.

Description

technical field [0001] The invention relates to the technical field of single target tracking, in particular to a deep target tracking method based on appearance enhancement. Background technique [0002] Target tracking has a very wide range of applications in the field of computer vision research, such as: tracking and focusing of cameras, automatic target tracking of drones, etc., all require target tracking technology. In addition, there are specific object tracking, such as human body tracking, vehicle tracking in traffic monitoring systems, face tracking and gesture tracking in intelligent interactive systems, etc. At present, target tracking mainly has two development directions: correlation filtering and deep learning. Especially since the introduction of various machine learning algorithms, the target tracking algorithm has shown a trend of flourishing. Algorithms based on deep learning have achieved extensive applications and breakthroughs in image and video reco...

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): G06T7/246G06K9/46G06K9/62
CPCG06T7/246G06F18/22
Inventor 王伟王侃苏航覃剑曾浩王欣
Owner 10TH RES INST OF CETC