DEVELOPMENT OF NORMALIZED INDICES FOR EXTRACTION BUILT-UP AREA BASED ON SPECTRAL CHARACTERISTIC OF WORLDVIEW-2 IMAGERY

摘要

Urban built-up is the impact of agricultural land and green open space conversion into settlements. Comprehensive data and effective mapping methods are needed in mapping dynamic land especially in urban areas. Mapping, planning, and monitoring of urban built-up area requires information about the land conversion. Mixing urban-built-up with the bare land, vegetation, and water reflection is a problem in land conversion using medium resolution imagery. Remote sensing images are useful for monitoring the spatial distribution and growth of urban built-up areas because they can provide timely and synoptic views of urban land cover. Remote sensing data, especially worldvicw-2 imagery, has advantages in both spatial and spectral resolutions. High spatial characteristics and combined with various spectral make the worldview-2 imagery is the best choice as the main data in the study for urban phenomena. The pure-band combination indices and PCA become the simple input for creating normalized index (NI). NI is a simple formula for index transformations. This study aims to explore the performance of the whole spectral information offered by the Worldview-2 sensor for built-up automatic extraction and to find the best accuracy of the normalized difference index produced by spectral reflection of worldview-2 images. The first step is to combine eight pure-indices images with normalized index. Then, make NI from eight PCA bands combined with pure-indices band of worldview-2. Visual analysis and interpretation is carried out by researchers to continue determining threshold values that distinguish built-up and non-built-up area. The number of combinations produced were 56 combinations of pure indices and 56 combinations of PCA. The results showed that the combination of pure-indices had better accuracy than PCA involving 8 bands at that time.