Method for describing and matching contour dense feature points in endoscopic image sequence

The invention provides a method for describing and matching contour dense feature points in an endoscopic image sequence, which comprises the following steps of: 1, enhancing an edge contour of a low-rank image obtained by adaptive RPCA decomposition on the basis of a bilateral filter, and extracting edge contour information on the basis of a gradient operator, obtaining the contour shape of the target through morphological operation; 2, extracting a multi-scale geometrical characteristic descriptor of a vertex on the contour shape to enable the multi-scale geometrical characteristic descriptor to be robust to the scale, the size and the rotation change of the contour shape; 3, performing dimensionality reduction on the dense feature descriptors of the contour shapes by using FFT, measuring the similarity degree between different contours by calculating the distance between the feature matrixes of the two contours, matching and tracking the target contour shape in the endoscope image sequence according to the