Dynamic Obstacle Detection in Traffic Environments

Abstract

The research on autonomous vehicles has grown increasingly with the advent of neural networks. Dynamic obstacle detection is a fundamental step for self-driving vehicles in traffic environments. This paper presents a comparison of state-of-art object detection techniques like Faster R-CNN, YOLO and SSD with 2D image data. The algorithms for detection in driving, must be reliable, robust and should have a real time performance. The three methods are trained and tested on PASCAL VOC 2007 and 2012 datasets and both qualitative and quantitative results are presented. SSD model can be seen as a trade-off for speed and small object detection. A novel method for object detection using 3D data (RGB and depth) is proposed. The proposed model incorporates two stage architecture modality for RGB and depth processing and later fused hierarchically. The model will be trained and tested on RGB-D dataset in the future.

A Survey of FPGA-Based Vision Systems for Autonomous Cars

“…A SSD network has two components the backbone, which is a pre-trained image classification network working as a feature extractor, and the head, which performs the final classification for subdivisions of the image. In [210], the backbone is based on a ResNet50.…”