Research on Vehicle Detection in Right Turn Blind Zone of Autonomous-rail Rapid Transit Based on Fisheye Image

The Autonomous-rail Rapid Transit (ART) significantly enhances residents’ travel and promotes the development of smart cities. However, given its unique left-side driving mode, there exists a potential safety hazard when making right turns. To address this challenge, this paper introduces an intelligent method using fisheye images for blind zone detection during ART’s right turns. Firstly, fisheye cameras are used to capture fisheye images, and the dual longitude method is employed for distortion correction. Secondly, a vehicle detection method based on Single Shot Multibox Detector (SSD) is proposed, which combines Squeeze-and-Excitation (SE) attention mechanism, Feature Pyramid Network (FPN) and Dilation Block (DB), called SFD-SSD. Ablation experiments show a 5.31% mean average precision (mAP) increase on BDD100k and 7.68% on VOC datasets compared to baselines. Finally, the proposed cross-longitude distance measurement method demonstrates an average detection error of 3% for forward distance and 9.8% for lateral distance, providing convenience for ART’s assisted driving.