A Research On The Metric Model For Remote Sensing Entropy And Qulity

Accompany with the development of remote sensing and computer technology, remote sensing imagery has been widely used in the environment protection, disaster prevention or control, weather forecast etc. Now Digital Orthophoto Map (DOM), which been primary produced with airboned and satellite imagery, has become one of the important production styles of national fundamental geographic information. For long time, lacking of the quantitative quality evaluation standard, the DOM quality is evaluated based on the person's eye vision, descripted with qualitative language. The report of DOM quality is subjective and limited, so it confine the application level of remote sensing imagery.Purpose of this research is collecting and analyzing the related meterial and results, elaborating on concept of remote sensing imagery entropy and quality measuring, advancing the practical measuring model of remote sensing imagery entropy and quality, including some related technical index.At the same time, for testing the model is manipulable and practicable, some related experiments are made.First, based on information and probability theory, the characteristic of remote sensing imagery and information resource are analyzed, and then the computing model of digital remote sensing imagery entropy using Markov information resource theory is advanced. In term of the people's eye vision rule, determining the correletive radius make the model been viable. Then.two kinds of imagery (one is the imagery covering the same ground area and being acquired with the different remote sensing platform , the other one is the imagery covering the different ground area and being acquired with the same remote sensing platform) are selected,and two experiments of measuring entropy and affecting imagery entropy by manipulatning imagery are made.Though the relationship between quality of manipulatning imagery and it's entropy in the same imagery isn't found, the following conclusion is educed:1) The digital remote sensing imagery belongs to the discrete information resource of having memory. In measuring information entropy, Markov entropy is more accurate than Shannon-Wiener entropy;2) In measuring information entropy, the appropriate value of the correlativeIVradius may be 13;3) After manipulating imagery, the imagery entropy is smaller than original imagery;4) Except for adjusting the imagery brightness and contrast synchronously, all manipulating don't make the distinct entropy variety.In research oh the measuring model of remote sensing imagery quality, based on fuzzy mathematics and probability theory, various quality charactestic (quality element) and spatial data quality evaluation procedures are analyzed. Then the model is advanced, that using fuzzy synthetically evaluated method.the quality elements composing with imager geometric accurate, imagery Modulation Transfer Function (MTF), utmost imagery resolution and quality of reference. Farther, various determinant functions are provided. Then, as the updating of the software of the digital surveying and mapping product audit (4D checker), DOM-checker was designed and developed. Through some contrastive experiment, following conclusions are:1) Using fuzzy synthetically evaluated method, composing quality elements with imager geometric accurate, imagery Modulation Transfer Function (MTF), utmost imagery resolution and quality of reference, may more veraciously evaluation imagery quality;2) For the different quality grade, one imagery having the different fall-into value. So the quality description may be more veracious and reasonable.In the general, this paper discusses and advances the measuring model of imagery entropy and quality. But there are still more problems that need to research deeply, such as the relativity of pixels, the vector of fuzzy weight, etc.