A comparative experimental study of interest point detectors and feature descriptors for visual navigation

Feature detection and matching is the first important module of visual navigation system. Therefore, how to select a pair of good feature detector and descriptor is a problem that visual navigation system must face. So far, there are many kinds of feature detectors and descriptors appearing in the literature, and the methods used are quite different. Although scholars try to make up for some shortcomings of existing algorithms, each feature detector or descriptor has its own advantages and disadvantages, but there is no perfect scheme to adapt to all applications. In this case, it is of great significance to evaluate the characteristics of feature detectors and descriptors.