Performance Assessment of Three Major Image Enhancement Techniques for Geospatial Data

Abstract

Image enhancement is one of the most important but difficult techniques for image processing researches. It is normally done to improve visual appearance and provide a better technique for future automated image processing. Sources of mages include satellite, photography and aerial photogrammetry that are used for geospatial data processing. These images suffer from poor contrast and noise. To use these images effectively, there is the need to enhance the contrast and remove the noise from the image to increase its quality. There are different techniques for image enhancement but this study focused on image interpolation. This multi-resolution technique is useful for variety of fields where fine and minor details are important. In this research, the Nearest Neighbor, Bilinear and Bicubic image interpolation algorithm were compared. Using the aforementioned techniques, two images were enhanced in order to compare their strengths and processing speed. The results of the algorithm of Nearest Neighbor had low computational time, low complexity of algorithm and poor image quality. On the other hand, the algorithms of Bilinear and Bicubic had average and high computational time, average and high complexity of algorithm and average and good image quality respectively.