Hard Example Mining Method for Visual Perception Based on Multi-sensor Fusion

Hard examples for visual perception can improve the performance of object detection effectively in autonomous driving scenes. However, they are dif ficult and hard to obtain in the real world. A hard example mining method for visual perception based on multisensory fusion was presented. The obstacle...

Full description

Saved in:
Bibliographic Details
Main Authors: Jun LIN, Wenbo PAN, Jun YOU, Yanghan XU
Format: Article
Language:zho
Published: Editorial Department of Electric Drive for Locomotives 2021-11-01
Series:机车电传动
Subjects:
Online Access:http://edl.csrzic.com/thesisDetails#10.13890/j.issn.1000-128x.2021.06.013
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849729308417851392
author Jun LIN
Wenbo PAN
Jun YOU
Yanghan XU
author_facet Jun LIN
Wenbo PAN
Jun YOU
Yanghan XU
author_sort Jun LIN
collection DOAJ
description Hard examples for visual perception can improve the performance of object detection effectively in autonomous driving scenes. However, they are dif ficult and hard to obtain in the real world. A hard example mining method for visual perception based on multisensory fusion was presented. The obstacles target segmented from radar point clouds was used to verify the image detection targets.Hard samples of image object detection and unlabeled samples in the real open world can be identi fied by the mapping relationship of multi-sensors based on actual obstacles. These hard samples were for training a new object detection model and remote deployment through the cloud-side collaboration mechanism to realize the optimization and iterative update of the model. Experiments show that hard samples in the mining autopilot scenes can be effectively collected by this method, and be used to improve the performance of object detection through incremental transfer learning signi ficantly. In addition, the algorithm also has important guiding signi ficance for autonomous driving scenarios in rail transit and other fields.
format Article
id doaj-art-79f6a471d84d47898a0286fad119c2c6
institution DOAJ
issn 1000-128X
language zho
publishDate 2021-11-01
publisher Editorial Department of Electric Drive for Locomotives
record_format Article
series 机车电传动
spelling doaj-art-79f6a471d84d47898a0286fad119c2c62025-08-20T03:09:15ZzhoEditorial Department of Electric Drive for Locomotives机车电传动1000-128X2021-11-010939922773180Hard Example Mining Method for Visual Perception Based on Multi-sensor FusionJun LINWenbo PANJun YOUYanghan XUHard examples for visual perception can improve the performance of object detection effectively in autonomous driving scenes. However, they are dif ficult and hard to obtain in the real world. A hard example mining method for visual perception based on multisensory fusion was presented. The obstacles target segmented from radar point clouds was used to verify the image detection targets.Hard samples of image object detection and unlabeled samples in the real open world can be identi fied by the mapping relationship of multi-sensors based on actual obstacles. These hard samples were for training a new object detection model and remote deployment through the cloud-side collaboration mechanism to realize the optimization and iterative update of the model. Experiments show that hard samples in the mining autopilot scenes can be effectively collected by this method, and be used to improve the performance of object detection through incremental transfer learning signi ficantly. In addition, the algorithm also has important guiding signi ficance for autonomous driving scenarios in rail transit and other fields.http://edl.csrzic.com/thesisDetails#10.13890/j.issn.1000-128x.2021.06.013hard example miningmulti-sensor fusiondeep learningvisual perceptionautonomous driving
spellingShingle Jun LIN
Wenbo PAN
Jun YOU
Yanghan XU
Hard Example Mining Method for Visual Perception Based on Multi-sensor Fusion
机车电传动
hard example mining
multi-sensor fusion
deep learning
visual perception
autonomous driving
title Hard Example Mining Method for Visual Perception Based on Multi-sensor Fusion
title_full Hard Example Mining Method for Visual Perception Based on Multi-sensor Fusion
title_fullStr Hard Example Mining Method for Visual Perception Based on Multi-sensor Fusion
title_full_unstemmed Hard Example Mining Method for Visual Perception Based on Multi-sensor Fusion
title_short Hard Example Mining Method for Visual Perception Based on Multi-sensor Fusion
title_sort hard example mining method for visual perception based on multi sensor fusion
topic hard example mining
multi-sensor fusion
deep learning
visual perception
autonomous driving
url http://edl.csrzic.com/thesisDetails#10.13890/j.issn.1000-128x.2021.06.013
work_keys_str_mv AT junlin hardexampleminingmethodforvisualperceptionbasedonmultisensorfusion
AT wenbopan hardexampleminingmethodforvisualperceptionbasedonmultisensorfusion
AT junyou hardexampleminingmethodforvisualperceptionbasedonmultisensorfusion
AT yanghanxu hardexampleminingmethodforvisualperceptionbasedonmultisensorfusion