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

Vehicle re-identification method for UAV platform based on multi-level attention mechanism

An attention, drone technology, applied in scene recognition, character and pattern recognition, computer parts and other directions, can solve the problems of insufficient clarity of the image to be recognized, complex changes in the perspective of the target object, etc., and achieve the effect of effective feature description

Active Publication Date: 2022-02-22
NORTHWESTERN POLYTECHNICAL UNIV
View PDF11 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] In order to solve the problem of insufficient clarity of the image to be recognized and complex changes in the viewing angle of the target object in the vehicle re-identification task for the UAV platform, the present invention provides a vehicle re-identification method based on a multi-level attention mechanism

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
  • Vehicle re-identification method for UAV platform based on multi-level attention mechanism
  • Vehicle re-identification method for UAV platform based on multi-level attention mechanism

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0019] Now in conjunction with embodiment, accompanying drawing, the present invention will be further described:

[0020] like figure 1 As shown, the main structure of the present invention can be divided into three branches. The global branch is a multi-objective task model whose backbone network is a ResNet-50 pre-trained on ImageNet. In this branch, the present invention takes full advantage of rich annotations by training multi-object models for re-identification, ID classification and attribute classification. In addition, in the bottom-up attention branch, the present invention uses a large amount of labeled discriminative region data to train a detector model as a discriminative region detector. The present invention utilizes the trained monitor to detect discriminative regions of the input image and construct a series of masks. By utilizing these mask values, the present invention uses mask average pooling to extract five local features belonging to specific discri...

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 relates to a vehicle re-identification method for an unmanned aerial vehicle platform based on a multi-level attention mechanism, and belongs to the field of vehicle re-identification. Use 3 three branches, global branch, bottom-up attention branch and top-down attention branch; the global branch outputs 2048-dimensional global features, and the bottom-up attention branch outputs 10240-dimensional Local features, the top-down attention branch outputs 2048-dimensional local features, and the features obtained by the three branches are spliced ​​together as the final feature output. Through the multi-level attention mechanism, the local features of the vehicle's discriminative area are effectively extracted, and the bottom-up attention mechanism is used to extract more effective vehicle global features. By combining the features under multiple attention mechanisms, the target instances for more efficient characterization.

Description

technical field [0001] The invention belongs to the field of vehicle re-identification, and is specifically a vehicle re-identification system combining top-down and bottom-up attention mechanisms. Background technique [0002] With the continuous development of UAV technology, vehicle retrieval based on UAV aerial photography can effectively help ground supervisors to search for designated vehicle targets. It has potential application value in many fields such as criminal investigation and traffic command. However, due to the influence of various factors such as UAV flight height, shooting angle, and light intensity, the vehicle re-identification task for UAV platforms has great complexity. [0003] Existing vehicle re-identification methods are often designed for surveillance video images, and usually rely on fine-grained features such as license plates and logos to assist in the recognition. However, in the vehicle re-recognition task based on UAV aerial images, due to ...

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): G06V20/58G06V10/56G06V10/764G06V10/82G06V10/74G06K9/62G06T3/40
CPCG06T3/4007G06V20/584G06V20/59G06V10/56G06F18/24133G06F18/22
Inventor 张艳宁张世周矫炳亮林蔚东邓玉岩
Owner NORTHWESTERN POLYTECHNICAL 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