Study On Target Detection Method Based On Information Fusion Of Automotive Radar And Camera

The domestic automotive market has grown faster in recent years,and the people’s enthusiasm for the purchase of automobiles and driving cars has been increasing too,but the attendant is the frequent traffic accidents,so the research and development of automotive safety technology are also rising in the automotive industry.Automobile Intellectualization becomes one of the key points,to make the automobile realize intelligent traveling or driving,we must realize the vehicle target effective detection.The effective detection of vehicle targets can provide the driver with target information to warn the driver of the road ahead,and also provide the basis for further automatic driving research.Based on the research and development of Enterprise’s automatic driving technology,this paper uses millimeter wave radar and camera sensors to gather information,realize the information fusion of target data,and verify the test through the ideal road information.The main research works are as follows:Firstly,this paper analyzes the development of sensor application and detection technology at home and abroad,based on the condition of our research,designs the general scheme of the basic vehicle target detection,and makes clear the connection of each part of the scheme.Secondly,using millimeter-wave radar sensors,the target of vehicle is primary chosen.The validity of the data information is verified based on the effective life cycle method.Using the symmetry,Haar-like characteristic and Adaboost learning method of vehicle target,accomplished the effective recognition of vehicle target on the basis of a lot of positive and negative sample learning,and the recognition result is optimized by Kalman filter,then realized the dynamic detection of vehicle target.Finally,the fusion model of millimeter wave radar sensor and camera sensor in time and space is established.Through the program and software,the actual equipment and road test are carried out,and the scheme and algorithm are verified.The results of the test show the validity of this method to measure the vehicle target.