A Sorting Method of Irregular Feature Point Array in Camera Calibration

Camera calibration is an important factor affecting the precision of vision measurement, and the correct ordering of target feature points is the precondition of high-precision vision measurement. Most of the existing methods of feature point array sorting have good sorting effects for regular arrays, but with further research on high-precision vision measurement, the target of irregular feature point array with higher precision of feature points extraction or anti-noise ability is proposed gradually, and the sorting requirement of these irregular feature point arrays is difficult to meet. In order to achieve the correct sorting of irregular feature point arrays, this paper proposes a method for sorting the feature point arrays based on convex hull and projection matrix. The convex hull and convexity are used to identify and locate the corner points of the feature point array. The projection matrix maps the feature point array column to a non-rotational and non-translation array, which is sorted according to the established rules to realize the automatic ordering of the irregular feature point array. The effectiveness of the proposed method is verified by comparative experiments.