Information content characterization in remote sensing imagery based on classification accuracy

The information content in remote sensing imagery depends upon various factors. Various textural measures are used to characterize the image information content. Our approach to quantifying image information content is based upon classification accuracy. We have developed a negative-exponential model that relates information content to spatial resolution, which is seen to be applicable to real images acquired by Landsat TM optical as well as SIR-C SAR sensors. An interesting conclusion that emerges is that although the TM image has higher information content than the SIR-C image at lower pixel sizes, the opposite is true at higher pixel sizes. The transition occurs at a pixel size of about 720 meters. This tells us that for applications that require higher resolutions (or smaller pixel sizes), the TM sensor is more useful for terrain classification. On the contrary, for applications involving lower resolutions (or larger pixel sizes), the SIR-C sensor has an advantage. Thus, the model is useful in comparing different sensor types for different applications