Multilevel Pipelined Processing for Aerial Image Restoration

Aerial images can get distorted due to various factors. During the acquisition of the images, it may get corrupted by distinct types of noises, camera misalignment etc. The restoration of degraded images has been a very important factor in the fields of medical, surveillance etc. Many algorithms in image processing do exist to restore degraded images but the efficiency of those in restoring the images is less. In this work, the aerial images are processed through various stages such as color stabilization, saturation enrichment and contrast improvement. In color stabilization, the color intensities (green, red, and blue) are globally adjusted. In saturation enrichment, to increase the saturation of an input aerial image, color of each pixel is projected onto bisectors of an green, red and blue color cube. And in contrast improvement, it is used to enrich the image quality by using the principle of unsharp masking. Real world aerial images were used to obtain the restored images. This work is verified for various input aerial images and the obtained restored images are verified qualitatively and quantitatively.