Abstract
This paper presents an outlook on the methods of stereo matching by computer which have been developed to date. The methods can be classified into
(1) Area correlation, (2) Matching of features and (3) Matching of segmented sub-areas.
In photogrammetry the area correlation method has been exclusively employed for stereo plotting. The other two methods have been developed in the field of ‘computer vision’. They are forecasted to be adaptable also in the photogrammetric field, especially for plotting large scale photographs, where occlusions occur frequently and objects with little textures show up.
Further the structures of some types of auto-correlators and their matching precision are briefly reviewed.
