Fuzzy classifier based enhanced Median filter for multi spectral Satellite images

Abstract

newlineImage segmentation is an essential component of the remote newlineSensing image inspection classification and pattern identification The road newlinesatellite image categorization points a momentous tool for the assessment of newlineimages In the present work the researchers have evaluated the computer newlinevision techniques for instance segmentation and knowledge based techniques newlinefor categorization of high resolution descriptions Segmentation and newlineClassification of patterns are an important area in a variety of fields including newlineartificial intelligence computer vision and image examination In such newlineproblems if a priori probabilities and the conditional probability density of all newlineclasses are unknown, then it produces worst results newlineFor sorting of the road satellite images, the technique named newlineAdapted Normalized Graph cut Segmentation with Boundary Fuzzy Classifier newlineObject Indexing AGSBFI is introduced Initially the satellite image is preprocessed newlineand is then segmented to have segments of barren land vegetation newlinearea building area and road area using enhanced median filtering method newlineThis method employs road extraction methods to determine the newlineimportance of removing the noise present in the satellite images for road newlinenetworks Moreover the probability of RGB values are converted to lab color newlinespace image for device dependency occurring at each position in the RGB is newlineemployed as the feature to precisely predict the lab color space image values newlinepresent in the satellite images of road networks newline newline