Enhancing remote sensing images by adjusting histogram globally and locally

Contrast enhancement is essentially needed in mapping preprocessing due to low average brightness and low contrast of original remote sensing images. The frequently used methods improve the visual quality of remote sensing images while compress gray ranges of skin-tone. As a result, shape and texture of these oversaturated regions are missing or distorted. The proposed method named GL-Enhancing not only enhance the contrast of remote sensing image, but also keep shape and texture. It is found that GL-Enhancing performs better than Equalization, Linear-Stretch, Gauss and SquareRoot methods in QuickBird 321bands true color image experiment. GL-Enhancing is efficient in mapping preprocessing. The proposed algorithm is easy to implement.