Low level vision processing and vegetation extraction using IRS-1C LISS III image

Removal of noise is essential for remote sensing applications, as its existence in images is inevitable and results in random misrepresentations to the region of interest, in the image. This paper proposes a method for low level vision processing and vegetation feature extraction for IRS-1C LISS III remote sensing images using a combination of partial differential equations (PDEs), and normalized band ratioing method in order pre-process and to extract vegetation features. The purpose of image low level vision processing was to remove the effect of noise, edge conservation, making the resulting segmented vegetation feature images approximate the ideal image. Effectiveness of the algorithm was proven statistically and visual interpretations of results are clearer. Results show that, the proposed algorithm obtains better performance on different frames of the study area with higher PSNR values, reduced RMSE and enhanced visual quality.