Research On Remote Sensing Image Quality Assessment In Remote Sensing Image Alteration Anomaly Extraction

Remote sensing technology has been used in various fields of the national economy, the products are widely used in the exploration of mineral, geographic information acquisition, geodetic survey, military investigation and agricultural research, etc. So remote sensing image quality problems related to the effects of product application, if not choose proper remote sensing data lead to the information provided by the conflicts with the actual geographical situation, its negative impact is huge. In the process of acquisition, transmission and storage of remote sensing image data, there are many factors that can reduce the quality of remote sensing image, so that the scope of application of remote sensing image is limited. In order to select the appropriate data from the mass of remote sensing data to produce and use, to improve the efficiency of the use of remote sensing image data, it is particularly important to evaluate the quality of remote sensing image.In this paper, based on the Landsat 8 OLI and ASTER remote sensing image as the research object, granular computing, rough set and fuzzy set and neural network theory are combined to establish a comprehensive evaluation model for remote sensing image quality assessment in remote sensing image alteration anomaly extraction. During the construction of index system in addition to the commonly used image evaluation index as the foundation, to study the methods of characterization of index and its corresponding characteristic parameter extraction method, and starting from the factors influencing the remote sensing alteration anomaly extraction, extraction parameters influencing factors, to build a geared to the needs of remote sensing alteration anomaly extraction of remote sensing image quality evaluation index system. Using ENVI remote sensing image processing software and MATLAB image processing tools and Rosetta rough set software, extracted the characterization of commonly used evaluation index parameters affecting the quality of remote sensing image, from the Angle of the influence of remote sensing alteration anomaly extraction, extraction factors influencing parameters, combined with the common index system, using fuzzy set theory to measure the discretization of continuous data processing, using rough set theory has been screened indicators, finally get a geared to the needs of remote sensing alteration anomaly extraction of remote sensing image quality evaluation of comprehensive evaluation index. At last, using the neural network theory, we can get the final result of the image quality.Taking Landsat 8 OLI remote sensing data as the experiment data sources to evaluate the accuracy of the model. The results showed that: the evaluatioin index system for remote sensing alteration anomaly extraction of remote sensing image quality evaluatioin built in this paper have better representation. By using the evaluation model constructed in this paper, the quality of the image can be reflected, and it has guiding significance on how to select the image data with good quality.