Fusion of Airborne LiDAR Data and Satellite SAR Data for Building Classification

In an airborne photogrammetry, a geometrical modeling and object classification can be automated using color images. Stereo matching is an essential technique to reconstruct 3D model from images. Although, object classification methods are automated using height data estimated with the stereo matching, it is difficult to recognize construction materials, such as woods and concrete. The construction materials are significant attribute data in building modeling and mapping. Therefore, ground survey and manual editing works are required in attribute data classification. In the LIDAR measurements, modeling and object classification are also automated by a segmentation of point cloud data. The intensity data also assist the object classification. Moreover, data fusion approaches are proposed using aerial images and LIDAR data. These approaches focus on modeling accuracy improvement and processing time improvement. However, these approaches classify geometrical attributes. On the other hand, although geometrical data extraction is difficult, SAR data have a possibility to automate the attribute data acquisition and classification. There are many researches related to monitoring activities of disaster, vegetation, and urban. Therefore, we focus on an integration of airborne LIDAR data and satellite SAR data for building extraction and classification. In this study, we use airborne LIDAR and satellite SAR data to classify buildings. Firstly, we generate a digital surface model (DSM) from point cloud acquired with airborne LIDAR. Secondary, the DSM is registered with a normalized radar cross section (NRCS) image calculated from SAR data. Thirdly, buildings are extracted from the DSM. Finally, the buildings are classified into several clusters using NRCS in the DSM. Although our result included noises such as bridges and automobiles, we classified buildings into clusters with average NRCS values.