A TCT-BASED METHOD FOR THE EXTRACTION OF MAN-MADE OBJECTS FROM IKONOS SATELLITE IMAGERY
碩士 === 國防大學中正理工學院 === 軍事工程研究所 === 96 === Man-made objects are usually the region of interest for image interpretator. That would assist searching targets in the interpretation work to classify man-made object classes with the multispectral information provided by remote sensing images (RSI) and then...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2008
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/50515961656713146393 |
| Summary: | 碩士 === 國防大學中正理工學院 === 軍事工程研究所 === 96 === Man-made objects are usually the region of interest for image interpretator. That would assist searching targets in the interpretation work to classify man-made object classes with the multispectral information provided by remote sensing images (RSI) and then to detect the edges of man-made objects in RSI. This research would stretch vegetation and man-made objects signals roughly by means of Tassaled Cap Transformation (TCT), extract targets of man-made objects with Self-Organizing Feature Map (SOFM) Network, and then detect the edges of man-made objects by Canny edge detection algorithm. We also adapt the Otsu optimal threshold method and K-means clustering to compare with SOFM. According to the result of this research, we would gain better edges through SOFM on the second stage of classification.
|
|---|