Based On The Pavement Of The Car Line Ccd Camera Data Acquisition And Processing

From the early 1990s, Foreign scientists began to study road information acquiring systems based on land-based mobile mapping technology, and established the road information database successfully, while Domestic research has just begun. With the domestic and international research focus - Mobile Mapping System, we designed and implemented the land vehicle based automatic road surface data acquiring system , which is consist of linear CCD camera, professional frame grabber board,, GPS (Global Position System), IMU (Inertial Measurement Unit)and other components. Then we discussed road line Recognition methods. The main contents are as follows:1) As road surface images are acquired during the vehicle moving on the roads, this paper discussed the reason of choosing linear CCD camera and methods and principles of selecting frame grabber board and other equipments.2) We designed the automatic road data acquiring system, and implemented the synchronization control program, Image-display program and Real-time images stored program.3) Image data processing conversion. After data conversion, the image data quantity is not only reduced by half, but the road information also has been well maintained.4) In the aspect of geometric correction, firstly, this paper described geometric deformation reasons. Secondly, according to linear CCD co-linearity equation, the geometric model and the geometric distortion correction model are set up. Finally, Simulation results show that geometric correction by co-linearity equation can compensate better than Polynomial correction, avoiding selection of ground control points. So the geometric correction can be finished on computer automatically.5) To find the best method to identify road markings, the maximum entropy segmentation, maximum classes square error segmentation and iterative segmentation are used to segment the typical road marking images. Experiment shows that using maximum entropy segmentation method you can identify road markings mostly accurately, but the effect is not very well in details as there is so much noise in images. Dynamic threshold method can identify the details of the images, but it caused some road misclassified into road markings. Finally, Dynamic threshold method integrated with the overall threshold method, this method can identify road markings accurately.