Implementation Of Lane Detection Based On Android Platform

With the rapid development of modern transportation industry and automobile industry, the frequency of road traffic accidents are becoming more and more obvious. It is significance to develop a kind of lane departure warning system that can predict a dangerous situation in advance and remind the driver to reduce traffic accidents. There is little development on lane departure warning system which is suitable for mobile devices. This paper mainly studies the implementation of the lane departure warning technology which is based on Android platform. It also increases the function of driving track playback for the system.Firstly, the structure of the system function is designed. It is based on two function modules. Lane detection module is implemented by JNI technology and track playback modules is implemented by Baidu Map API.Secondly, it uses Android camera to capture the road image and transfer them to the entrance of local image processing method which is generated by NDK. Through scaling the gray of the image, thresholding and setting the interest region of the image, it can eliminate the useless information and improve the real-time performance of the system. It uses the improved Sobel algorithm to detect lane edge and obtains the lane line parameters by Hough transform line detection, and displays it in the road image actually.Then, it obtains travel information by positioning system of Baidu SDK and uses SQLite database for data storage. Combined with Baidu Map API, it draw trajectory on the map according to the path information. It also shows whether the vehicle deviates from the lane and related information from trajectory.Finally, it tests the function of the system. And the results show that the system can realize the detection of the lane and accurately draw the historical trajectory exactly.