Based On High Resolution Remote Sensing Of The Chengdu Plain Rural Settlement Information Extraction Research

The resident of village, also named as"settlements", is the product of the interaction between human and nature and the basic premises of a variety of human activities. There is strong space correlativity between the space distributing of residential place and that of regional population,economy,property. It provides basic framework for the special extensity of statistic data for social economy such as population, economy, property, etc.. The development of remote sensing technology provides advanced means of the acquirement of the information of residential area. Acquiring the information for rural residents timely and accurately has an important guiding significance for construction and development of rural areas.In this paper, the technology of object-oriented remote sensing information extraction is taken, The basic prime is: First, it produced homogeneous image objects through segmentation technology, then analysed the characteristics of the object and extracted the object feature information from it. Optimizing the characteristics and establishing the database of the object feature information, it carried out classify by member function according to the object feature information,Finally, it extracted our goal according to classification result and the goal feature information. In this paper, it based on the technology of object-oriented remote sensing information extraction, for Complex and changeable for rural features, it will exist defects of over-segmentation or insufficient-segmentation for the traditional segmentation method, it research the key technology of extract rural residential area of high-resolution remote sensing image based on object-oriented technology. The main work include: different segmentation scales are used for different object to achieve better segmentation. At the same time, While doing feature extraction and knowledge discovery, not only the knowledge of spectral be used, but object shape, texture, and spatial relationships also combined, Finally identified and extracted units and components of rural settlements by using optimized remote sensing characteristics. At last, the Quickbird-0.61m high resolution remote sensing image in XinDu of ChengDu Plain as an example, for experimental verification. The experimental results and the accuracy of the evaluation results show that the precision of most components in settlements is more than 80%, the overall accuracy of settlement units is more than 90%, it shows that it can extract the component and unit information of resident by our proposed method in this paper.