An Accuracy Adjustment by Fusion Method with Gis Data and Remote Sensing Data

The Department of Construction in Kochi prefecture, Japan established and shared a bridge database without accuracy information. When the data overlay with a commonly used GIS data, some errors appeared. To adjust the accuracy of the bridge database, the intersections of roads and rivers of the National Land Digital Information (NLDI) of Japan were used in the accuracy adjustment, and 64% of bridges over rivers obtained accuracy adjustment (Jong Hyeok JEONG, Masataka TAKAGI, 2001). In the adjustment, 36% bridges could not be adjusted because of the limitation of the NLDI. The NLDI did not have enough information for the adjustment. To cover the limitation of the NLDI in the accuracy adjustment of the bridge database, an IKONOS satellite image was used. Expected bridges were extracted from an ISO unsupervised classification of the IKONOS image and edge detection of the IKONOS image, and then the extracted bridges were converted to polygons. Each bridge polygon in the bridge database matched with the nearest extracted bridge point to make a one-to-one relationship. Then the attribute data of each bridge in the bridge database added to the coordinate of the nearest polygon. In the test area of this study, 23 bridge points existed, 15 bridge points of them were adjusted by the extracted bridge position from classified image, 19 bridge points of them were adjusted by the extracted bridge position from the edge detection of the IKONOS image. In this study, the edge detection method provided reliable result of accuracy adjustment.