A SUB-PIXEL VISUALIZATION METHOD TO DISPLAY FUZZY PHENOMENA USING RGB COLOR COMPOSITE (CASE STUDY: MANGROVES FOREST)

Abstract

Natural phenomena boundaries and complexity of features in an urban area due to the low spatial resolution, lead to more pixels of satellite images included in reflectance of multiple land-cover/object components. The sub-pixel information extracting model outputs are fractional cover maps of interested class (end-member), with membership values between zero and one. These maps represented gradient change in only one fuzzy phenomenon boundaries such as vegetation cover. However, in multiple fuzzy class area or complex fuzzy phenomena such as mangrove forests, in the northwest of the Qeshm Island, Hormozgan, Iran, displaying several fractional covers may cause confusion and misunderstanding for the end-user. In this study, an additive color composite and spectral mixture analysis method is utilized for multiple fractional cover representation. The proposed method is implemented on images acquired from Operational Land Imager (OLI) sensor in the Landsat 8 satellite to extract three fractional covers (water, vegetation, and soil). An RGB color composite was used for each type and percentage of fractional cover for given pixel to display fractional cover separately. Based on such RGB color composite represented both quantitative and qualitative information, we used the RGB color solid cube as map legend for better understanding and map interpretation. The result of this study showed that suggested sub-pixel visualization method, gives new vision to the end-user understanding of fuzzy phenomena.