Evaluation of Perceptual Contrast and Sharpness Measures for Meteorological Satellite Images

Abstract

Sharpness and contrast have great impact on perceived quality of an image. This paper focuses on sharpness and contrast measures to evaluate quality of Thermal Infrared (TIR1) channel of Indian National Satellite-3D (INSAT-3D) without using any reference image. Most of the sharpness metrics can scarcely manage to discern image quality degradation against high frequency behavior due to noise. Six Image Quality Measures (IQMs) are employed to study their behavior in terms of blur, noise and intensity changes simultaneously. Results show that (1) change in value of Measure Of Enhancement By Entropy (EMEE) is more discernible with change in contrast of an INSAT-3D image as compared to other measures and (2) Second Derivative Like Measure Of Enhancement (SDME) has a potential to distinguish high frequency content due to sharpness arisen due to un estimated noise up to some remarkable level in case of TIR1 INSAT-3D satellite images. Performance comparison of six measures against blur, noise, contrast and sharpness changes is presented.