Quality Evaluation Of High-resolution Image Based On Multi-level Fuzzy Comprehensive Evaluation

Along with the launching of China s GF-1 satellite, high-resolution image has developed the main source of image data in China, its quality will have significant influence on the data exchange, sharing and use, but also will directly impact on the accuracy and reliability of GIS application,analysis, decision-making. Therefore the majority of users pay more and more attention to the high-resolution image quality. Evaluating high-resolution image quality sensibly, objectively, comprehensively and fairly is vital to distinguish the quality grade of the different image, as well as to ensure High-resolution image thematic products serving national economic development and national defense construction effectively. Aiming at the quality evaluation of high-resolution image, the main research work is as follows:1. Construction of high-resolution image quality evaluation index system. The two-layer fuzzy comprehensive evaluation method is introduced into the comprehensive evaluation of the high-resolution image quality. Taking GF-1 satellite image as typical case study, through the analysis of existing quality evaluation index system at home and abroad,and combined with the characteristics of our country's high score image data, constructing the index system of high-resolution image quality, it includes grid quality, mathematical accuracy, logical consistency, attribute accuracy, attachment quality and characterization quality, six first-level indexes and 23 second-level indexes, which provides a basis for high-resolution image quality evaluation analysis.2. Construction of multi-level fuzzy comprehensive evaluation model.Due to the traditional methods have many shortcomings for the fuzzy attributes of images, we cannot make a full and reasonable evaluations on them. This paper proposes the theory and builds the model about the multi-level fuzzy comprehensive evaluation method. The weight and membership of the factor set are analyzed quantitatively based on analytic hierarchy process and triangular distribution membership function. It overcomes the random and fuzzy uncertainties of spatial data itself and the lack of quantitative analysis in existing research.3. Example analysis of high-resolution image quality evaluation. This article relies on high-resolution city meticulous management remote sensing application demonstration system project. Multi-level fuzzy comprehensive evaluation model is used to analyze the image quality of small town of Huoshan county and Huangshan scenic area, and it is concluded that the ratings of the high score image quality are "good" and the evaluation results are basically in accordance with the actual situation of the project.It is more scientific to use fuzzy synthesis evaluation on High-resolution image. The method can not only accurately analyze the image quality of the plain and hilly area, but also has a good division of the image quality of the mountain area which has good use value.