Information Entropy of Remote Sensing Images and its Applications in Image Fusion

摘要

Aiming at improving radiometric resolution in remote sensing images of the new generation, this paper developed a method for double byte joint entropy based on proper quantization, which is known to affect entropy, and decorrelation of pixels and bands in multi-spectral images. Joint entropy for images is computed and analyzed, which is close to non-memory, implying that the result is more accurate. The proposed method can be used for the optimum bands selection and quality evaluation in image fusion, which is valuable for increasing precision in entropy analysis.