In an automated procedure for stereoscopic mapping, reasoning is always needed when insufficient information due to occlusions, shadows, or indistinctness of image features is encountered. In order to alleviate the influence of the information insufficiency, it is common to include more geometric or spectral data source to assist the reconstruction for 3-D man-made buildings. Considering the improvement of geometric information, this investigation uses aeria1 quadruplets with 60% overlap in two perpendicular directions to perform the reconstruction for man- made buildings. Two major components include (1) generation of digital surface mode1 (DSM) by templet matching, and (2) feature extraction and processing such as image smoothing, edge detection, edge thinning, and CLR matching. Based on the blocks extraction from DSM, each potentia1 building is extracted followed by edge detection, Hough transform to detect straight lines, line grouping, and 3-D positioning for bui1ding corners. All the stereopairs composed by 4 images are integrated in this study. Experimental results indicate that, as compared to stereopairs quadruplets cou1d be extracted more edge information for better 3-D building reconstruction . In addition, it is also possible to complete the gable roof when only partial edge information is available by using dual searching between image space and object space.

Three dimensional geographic data is getting important. For those applications such as 3-D city model, landscape visualization, selection of base station for telecommunication, the topographic data is indispensable. Photogrammetric mapping technology is still the major approach for 3-D topographic mapping. The technology has evolved from analogue to analytical one and now toward automation. In the automation, aerial triangulation has been successful and in operation. On the othe hand, the automated determination for ground object is still in the early stage. Reconstruction of man-made building, in particular, is the major task in the photogrammetric automation. On the contrary to natural objects, man-made buildings are with regular boundaries and elevation drops. Thus, it is needed to extract corner points, edges, regions, and elevation distribution in reconstruction for man-made buildings. This investigation uses aerial color stereoscopic images to reconstruct those buildings. Considering spectral and geometric characteristics, the components of the reconstruction are：(l)color transformation, (2)region growing, (3)edge detection, and (4)corner detection. Finally, we develop a CLR matching algorithm which uses three windows to determine and to validate the disparities for the building boundaries.

Clustering algorithms are commonly used to perform an unsupervised classification, to expose internal data structures, and to facilitate the selection of training samples for a supervised classification. The iterative migrating means clustering algorithm (IMMCA) has several input parameters, some of them are difficult to determine, and thus clustering results are easily affected by inappropriate values of input parameters. Moreover, IMMCA is inefficient due to its iterative clustering process. The objective of this study was to develop a hierarchical partitioning clustering algorithm which can alleviate the problems stated above and serve as an alternative to IMMCA. The clustering algorithm was developed based on the concepts of hyperplane and binary tree. The two clustering algorithms were compared in the unsupervised classification of the video image data in terms of friendliness, efficiency, influence of input parameters, and classification accuracy. The new algorithm needed to determine only an input parameter, the maximum number of cluster, and was users friendly. It was much more efficient than IMMCA, but had a slightly lower classification accuracy than that of IMMCA. Therefore, it could be used as an alternative to IMMCA when applied to an unsupervised classification of land cover.

This study is to integrate and analyze the impact on water conservation due to deforestation or other land use pattern changes. Geographic Information Systems (GIS) were adopted to build a digital database for reservoir area and hydrologic analysis. Remote Sensing (RS) and Global Positioning System (GPS) were performed to extract and identify the surface vegetation details and land use changes information. Hydrologic model (Hydrological Simulation Program­Fortran； HSPF) for water conservation capacity estimation from land use changes at Da-Pu Reservoir Watershed. This study found the major land use changes existed in watershed area were orchard and forest, but both change rates were below 5.5%. The outcome also demonstrated the effectiveness of simulating the relationship between heavy rainfall and runoff by applying HSPF model. This study can provide an important reference for sound watershed management.