OBSTACLES DETECTION BASED ON MILLIMETRE-WAVE RADAR AND IMAGE FUSION TECHNIQUES

摘要

An obstacles detection method based on the date clustering and information fusion of millimetre-wave radar and the camera in real traffic scenarios is proposed in this paper. The ROI (region of interest) established from millimetre-wave radar through data processing can be transformed into the image coordinate by multiplying perspective transform matrix derived from the co-calibration between millimetre-wave radar and the camera. The effectiveness and the robustness of the proposed obstacles detection algorithm are checked and identified by using PreScan traffic simulation software. The data processing and information fusion algorithm from millimetre-wave radar and the camera has been tested in laboratory of Chang An automotive engineering institude. The results show that the proposed date processing and information fusion algorithm can catch up the feature information of obstacles rapidly and accurately. The temporal-spatial consistency and the multi-level fusion algorithms in radar-vision fusion investigated in this paper supply an important technique tool to complete obstacles identification accurately and three dimensional environment recognition.