A Lane Departure Warning System Based On Android Platform Of Research And Implementation

With the rapid development of vehicles and residents holding auto industry continues to grow, the car has become the main means of transport for people to travel. For safety travel and reduce the number of traffic accidents, Automotive safety technology to become the focus of the study. Automotive safety technology into active and passive safety technology, Active safety technology for early warning before the accident, and Passive safety technology is used to reduce the damage caused by accident. Lane departure warning is an important part of automotive active safety technology, while driving the vehicle, the driver has no intention to transform lane of the vehicle from the original lane, the system will warn him by sound, vibration or other form of warning, it is possible to effectively reduce the probability of accidents. LDWS includes lane marking line’s recognition, lane marking line’s tracking, lane departure warning and identification. This paper studies and discusses the lane departure warning system implementation, and transplant system to Android device platforms. Research work and results mainly include the following aspects:1. Obtain the road image through the Android mobile phone’s camera, using unidirectional gradient detection operator, OTSU algorithm and local maxima of nonfiction to extract the lane line edges, using improved Hough transform to extract linear parameters, and extract each lane marking line inner edge line.2. Using lane edge line parameters of and θ to construct a linear tracking model. In order to eliminate noise interference, using Kalman filter to track the lane marking line’s peak position in the space of Hough transform.3. Because we can’t obtain the intrinsic and distortion parameters of camera easily, and get the homography matrix of inverse perspective transformation, we use a simple warning method to estimate the vehicle position in the lane.4. Transplant the lane departure warning system to Android device platform. Through Android’s Java API for image capture, video frame display, interaction with user and feedback from LDWS and alarm functions; The lane departure warning algorithm is compiled into a dynamic shared libraries with NDK; use JNI to call the system algorithm processing pavement frame, according to the identification results, if the displacement occurrence through the loudspeaker warning.