Contour Extraction Method of Stereo Images Using Sensor Model and Image Matching

摘要

The present invention extracts the contours directly from the stereo image by using the image matching method that assumes the transformation relationship between the sensor model established in the stereo image and the corresponding pair of pairs. To provide a contour extraction method of stereo images using model and image registration.;The contour extraction method of the stereo image of the present invention establishes a sensor model from the received stereo images and performs image registration on the stereo images. The contour extraction method of such a stereo image includes a first step (S203) of finding a corresponding point pair corresponding to the same point of the stereo image, and a point in each adjacent image having the same height value as that of the corresponding point pair of the first step (S203). A second step (S204) of estimating the location of the field. In addition, the contour extraction method of the stereo image includes a third step (S205) of matching the corresponding pairs of the points estimated in the second step (S204), the second step (S204) and the third step (S205). Iterating (S206) to include a fourth step (S207) for obtaining a series of matching points having the same height value. In addition, the contour extraction method of the stereo image includes a fifth step (S208) of extracting the contour lines from the matching points of the fourth step (S207).