Comparison of feature points detection methods

Abstract: Methods of selecting feature points are widely used in tasks of comparing images, combining panoramic images, object recognition. Currently, a large number of methods have been developed that detect feature points. Among them are methods proposed by such authors as H. Moravec, C.G. Harris, J. Shi, C. Tomasi, W. Förstner, S.M. Smith, J.M. Brady, E. Rosten, T. Drummond and others. Methods that successfully detect feature points on images in one subject area often show lower results on images from another subject area. This paper compares the best-known methods of selecting feature points.

Keywords: industrial robot, computer vision, image processing, feature points