| Summary: | Carrying out a proper the development rural areas in small scales requires accurate and updated geo-spatial data and information, which are traceable by using Global Positioning System (GPS) navigation instruments. This research is intended to recover the positions of buildings and fishponds processes, wiped out by the 2004 tsunami, Blang Oi Village by using GUI on Matlab.
The study was conducted in Blang Oi Village based on the IKONOS satellite images at pre and post-tsunami times, and the QuickBird satellite image after reconstruction with the scale of 1:2,000, 436 coordinate points obtained with GPS Garmin RING 530HCx hand help receiver. Geometric corrections ware performed on the three satellite images by placing four control points on the individual image for conformation the Sobel operation to differentiate object and background. Was used morphological operations ware exploited to connect the disconnected boundary lines by the Sobel operation, and with mean, median and Wiener filter on 3x3 windows to minimize noises in the image. The corrected GIS was developed on segmentated, rastered, and vectored images by applying UTM zone coordinate system 46N Dantum WGS 1984.
The location of construction object and the lost pond borders were detected by overlaying segmentation image for pre-tsunami and by using raster image at post-tsunami times. The differences of construction object positions developed were indicated by overlaying segmentation image pre-tsunami by using raster image after the reconstruction times. The GPS data were used to check the positions of construction coordinate points on the Google Earth location data. The results show the re-constructed objects were not accurately detected, while for the ponds object the accuracies are superficies good detected. Manually digitalization results for the pond objects through the use of ArcGis was similar to the result of thresholding 40-based Sobel operation but different for building objects.
|
|---|