A Map-Based Approach to Extracting Object Information from Aerial Images

We have developed a map-based approach which enables us to efficiently extract object features from aerial images. We match an image with i ts corresponchng map in order to estimate the object states in the image. This approach is characterized by i t s active use of a figure contained in the map a s the object model for comparison. We determine the principal steps for the map-based approach for recognizing objects, and the steps are applied for obtaining the locations of missing buildings and the heights of existing buildings. The results of experiments using aerial images of Kobe City show that the approach is effective for automatically extracting urban-object information from aerial images.