Feature Level Fusion of Landsat8 and Wordview2 Images for Improving Vegetation Detection in Urban Areas Based on Knowledge Based Method

Given the diversity of urban problems and the sensitivity of each of the specified spectral range of the electromagnetic spectrum, various imaging sensors exhibit different behaviors in relation to some of the toll record. The color distortion and respond to abnormal sensor spectral range identified as one of the most important challenges the correct diagnosis in Remote sensing, effects, and advocacy. Data integration in order to take advantage of spectral data recorded by the various sensors in better detection vegetation is also significant benefits. In this study, the results of two Landsat 8 and WorldView-2 images merge in order to better identify areas of vegetation in the urban area, using statistical transfer PCA is used. Among the benefits that this integration as well as integration and coherence spectral data cited two sensors. Using Landsat 8-band SWIR image, increase transparency and thus improve detection and associated vegetation. In this paper, after the merger of two images taken from an urban area in Tehran at the level of spectral characteristics, process knowledge base to identify and classify areas of vegetation was carried out which resulted in the mining areas of vegetation with an accuracy of 81.3% was achieved.