THE CHANGE DETECTION OF BUILDING MODELS USING EPOCHS OF TERRESTRIAL POINT CLOUDS

The highly detailed building modeling through terrestrial laser scanning has been studied in recent years. Among the applications of highly detailed models, change detection from street level, which has small scale, is relatively new and has good potential. Hence based on the rebuilt building models, we propose an approach of change detection of building models deploying different epochs of TLS data, which follows three steps: automatic point cloud registration, building model change detection and quantification of changed regions. The presented method only focuses on the disappearing change (for instance erosion, damage, etc.) of rebuilt building models and quantification of the changed regions, which can be helpful to the works, e. g. disaster management, insurance claim evaluation and so forth. Since the accuracy of terrestrial laser scanning is normally in the order of millimeter, the accuracy of the computed planar surface area is expected to reach the order of centimeter taking the error accumulation effect into consideration, which certainly can meet the need of the practical applications.