An Efficient Wide-Range Pseudo-3D Vehicle Detection Using A Single Camera

Wide field-of-view and distant (wide-range) vehicle detection enables active safety in intelligent driving systems. However, existing vehicle detection methods based on rectangular bounding boxes (BBox) often struggle with perceiving wide-range objects, especially small objects at long distances. And BBox expression cannot provide detailed vehicle geometric shape and pose information. This paper proposes a novel wide-range Pseudo-3D Vehicle Detection method based on images from a single camera and incorporates efficient learning methods. This model takes a spliced image as input, obtained by combining two sub-window images from a high-resolution image. This image format maximizes the utilization of limited image resolution to retain essential information about wide-range vehicle objects and effectively improves the detection performance of small targets. To detect pseudo-3D objects, our model adopts specifically designed detection heads. These heads simultaneously output expanded BBox and Side Projection Line (SPL) representations, which capture vehicle shapes and poses, enabling high-precision detection. Furthermore, a joint constraint loss combining the object box and SPL is designed to enhance the detection performance. Experimental results on self-built and KITTI datasets both demonstrate that our model achieves favorable performance in wide-range pseudo-3D vehicle detection across multiple evaluation metrics. Our demo video is available at https://www.youtube.com/watch?v=1gk1PmsQ5Q8.