| This thesis is mainly based on the work of Anhui Province Science and Technology Research Program (08010204253): "embedded drive safety system with visual capability" in 2008 and Anhui Technology Project 2008: (08020203013) "Video-based multi-source image fusion splicing technology research".The projects aimed drive safety problem, and given an "embedded drive safty system with visual capability." The key problem of this system is video image mosaic algorithm. Against this problem, this thesis provides a solution, a video image mosaic algorithm based on Template matching by the adaptive. This thesis includes the following contents:(1) This paper puts forward a video image mosaic algorithm based on template matching by the adaptive, this algorithm includes: geometric correction, corner detection, template scanning, template matching, best seam-line scanning and image fusion.(2) Design of the camera calibration algorithm by studying on camera model, geometric correction via camera intrinsic parameters.(3) Design of the adaptive template matching algorithm based on corner detection and voting mechanism, finding the best seam-line to stitch images based on dynamic programming, effectively solving ghost difference between adjacent motion objects in overlapping area.(4) Images are smoothed by gradually fade in-out theory after mosaics of images, and both accuracy analysis and real-time analysis. |
PDF Full Text Request