Extraction of Linear Features by Edge Detection Techniques

The term ―feature‖ in Remote Sensing Image (RSI) takes its specific definition from the objective and scope of the study by the analyst. In spatial data mining using remote sensing satellite data, features mainly indicate objects constituting natural resources such as land, water and sea. This may broadly encompass vegetation, land condition, water quality, extent and types of vegetation and combinations these features. Extracting features in spectral domain is a tedious task requiring combination of image processing techniques such as Principal Component Analysis (PCA), statistical operators and clustering methods. The PCA as discussed in the chapter 4 has helped to understand the necessity of adapting certain mathematical functions not only to reduce the multidimensional image but enhancing the image to identify and extract relevant features. Similarly, boundary detection and detecting linear features from RSI involves enhancement techniques using statistical operators.