基于云理论的遥感图像土地利用分割分类方法研究

传统的遥感图像土地利用分类技术自动化和智能化的程度较低,遥感图像土地利用分类问题是一个包含随机性和模糊性的不确定性问题,而云模型把模糊性和随机性集成到一起,构成定性和定量相互间的映射。据此,本文探索将云理论引入到遥感图像土地利用分割分类技术中,建立了基于灰度的云映射空间,实现对遥感图像的分类。同时以武汉市南湖地区进行了实证分析,进一步阐述了模型构建过程,通过分类结果的评价和对比,探索该方法在本领域中的适用性。

Landuse Classification and Evaluation of Remote Sensing Image Based on Cloud Theory

The traditional landuse classification technology of remote sensing image is in a low level of automation and intelligence.The landuse classification of remote sensing image is an uncertain problem which contains random and fuzziness.The cloud model integrates fuzziness and random together,constituting mapping between qualitative and quantitative.According to the above,this article introduces the cloud theory,establishing a cloud mapping space based on gradation,aiming to solve the problem of landuse classification of remote sensing image.At the same time,this paper has constituted the result evaluating indicator system in aspect of the homogeneity indexes of the region and the boundary location,and has carried on the empirical analysis of the SPOT image of Nanhu area of Wuhan,elaborated the model construction process in depth,explored the serviceability of the method by evaluation and contrast of the classification results.