A navigation method based on monocular video path is proposed on the basis of double-view localization in monocular vision. The video path is expressed by a set of ordered key images extracted from monocular camera. The image feature points are extracted by adopting Harris corner algorithm

and the zero-mean normalized cross-correlation algorithm is used for image matching. According to constraints relationship between the matching points in double view based on monocular camera

their basic matrix can be obtained and optimized

and the pose parameters of the points can be derived

localization of mobile robot can be realized

and it can be navigated under the guidance of video path. The proposed navigation method is verified

which proves that it has good practicability in outdoor environment.