Measurement of Graylevel and Texture Information in Multispectral Images

A method of estimating the information in remotely sensed images based on the parametric entropy model is proposed in this paper. The goal was to quantify the information content of gray images, texture images, and combined images. The key point of this method is that u DPCM (differential pulse code modulation) can reduce redundant information caused by spatial and spectral correlation, can estimate noise preliminarily to avoid the increase of the entropy because of noise, and estimate the information content of images by making use of parameter-entropy model. Results from information assessment of TM data are reported and discussed; useful information content from multispectral remotely sensed images can be effectively assessed by the method described in this paper. Some application issues; such as classification or quality assessment of remotely sensed images, can be guided by the method.