{"title":"A Multi-scale Fusion Obstacle Detection Algorithm for Autonomous\u0000 Driving Based on Camera and Radar","authors":"Sihuang He, Chen Lin, Z. Hu","doi":"10.4271/12-06-03-0022","DOIUrl":"https://doi.org/10.4271/12-06-03-0022","url":null,"abstract":"Effective circumstance perception technology is the prerequisite for the\u0000 successful application of autonomous driving, especially the detection\u0000 technology of traffic objects that affects other tasks such as driving decisions\u0000 and motion execution in autonomous vehicles. However, recent studies show that a\u0000 single sensor cannot perceive the surrounding environment stably and effectively\u0000 in complex circumstances. In the article, we propose a multi-scale feature\u0000 fusion framework that exploits a dual backbone network to extract camera and\u0000 radar feature maps and performs feature fusion on three different feature scales\u0000 using a new fusion module. In addition, we introduce a new generation mechanism\u0000 of radar projection images and relabel the nuScenes dataset since there is no\u0000 other suitable autonomous driving dataset for model training and testing. The\u0000 experimental results show that the fusion models achieve superior accuracy over\u0000 visual image-based models on the evaluation criteria of PASCAL visual object\u0000 classes (VOC) and Common Objects in Context (COCO), about 2% over the baseline\u0000 model (YOLOX).","PeriodicalId":417451,"journal":{"name":"SAE International Journal of Connected and Automated Vehicles","volume":"99 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127125101","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}