A Vision-based Safety Driver Assistance System for Motorcycles on a Smartphone

In this study, we develop a vision-based safety driver assistance system (DAS) for use with motorcycles. DASs have been studied by researchers and developed by car companies as a tool to avoid traffic accidents. However, existing studies have seldom focused on motorcycle riders. In Taiwan, the prevention of motorcycle accidents cannot be ignored when developing safety DASs since they are involved in a significant percentage of traffic accidents. This paper proposes a safety DAS for motorcycles which can obtain an input sequence from a smartphone, detect the distance between the owner's motorcycle and an obstacle ahead, and output warning messages if necessary. The study is divided into two parts. Firstly, transformation matrices of different camera angles are calculated and stored in a transformation matrix database. Secondly, a kernel system is developed. The system can turn each input frame into a top-view image using a suitable transformation matrix from the database. Further, the system can detect obstacles in the top view image and calculate the distance between the owner's motorcycle and the obstacles. Concurrently, the GPS embedded in the smartphone can be used to obtain the speed of the motorcycle in order to calculate the safe distance. Finally, the system gives the rider suitable warning messages after comparing these two distances. The experimental results show that the proposed real-time system is robust and efficient. Moreover, the system is also suitable for deployment in cars.