A FAST METHOD WITH SEARCHING WINDOW CONSTRAINT FOR MOVING OBJECT DETECTION ON A MOVING PLATFORM

摘要

The study we propose is about moving object detection on a moving platform adopting the merits of the triangular geometry, the color segmentation, the disparity map, and the compensation of the background using the two dimensional plane transformation. Therefore, an image processing in real circumstance is realized in a short computational time. Because the algorithm is not complex, the moving objects can be detected soon with high accuracy. In addition, it is robust to reconstruct the complex view with various depths of fields.