Study Of Uncertainty And Computating Methods Of Information Quantity Of Remote Sensing Image

Information entropy is used which is based on information theory. Via the computation about Remote sensing image, put the uncertainty problem about Remote sensing image up to a height of information theory and study it, and found the consolidate mathematics expression between information quantity and uncertainty about t Remote sensing image in order to resolve the quantitative evaluation problem between information quantity and uncertainty of Remote sensing image.The uncertainty of spatial data is analyzed systemically as well as the uncertainty of Remote sensing data; the concept of uncertainty, uncertainty and error and the connection with each other are analyzed contrastively; new research method is brought forward which studies the uncertainty of Remote Sensing image form the point of view of information theory, and the meaning of uncertainty and uncertainty are defined newly which is uncertainty indicates information, but uncertainty is a measurement index for uncertainty which indicates information quantity included in uncertainty which is completely eliminated; At last, the computation methods and flow are given and a lot of experiments are carried out.TM image, SPOT image and IKONOS image are selected as main data sources; exert ERDAS IMAGING software to process Remote sensing image's subset and transforming; exert MATLAB software as a flat roof and compile M-file procedure to compute the information quantity of Remote sensing image.Firstly, four different areas of interest are selected with 231 row and 231 column in the same TM image which is city, water, cropland and forest, and data quantity is computed separately as well as information entropy, noise uncertainty, self-correlation coefficient of neighboring pixels, mutual information quantity and average information quantity, bands correlation and practical information quantity. Secondly, TM data is selected with three different time of Nenjiang area, and information quantity of Remote Sensing image is computed. At last TM image, SPOT image and IKONOS image are selected as data source, and interest areas of same size are selected to be computed information quantity of Remote Sensing image with different spatial resolution.The computation results show the fact that information quantity of Remote sensing image can be affected by the intensity quantity scale, color, noise,...