Font Size: a A A

Research On Object-oriented Extraction Method Of Impervious Surface Over Mining Area

Posted on:2013-07-12Degree:MasterType:Thesis
Country:ChinaCandidate:X T GeFull Text:PDF
GTID:2230330392950259Subject:Surveying the science and technology
Abstract/Summary:PDF Full Text Request
At present, with the development of the national economy and society, the miningactivities is increasingly growing, which causes the destruction of land resources andrapid land coverage changes in mining area. The mining area is facing seriousenvironmental problems. Extraction of mine features is the precondition and foundationof disaster monitoring and environmental assessment. Thereinto, the impervious surfaceis one of the typical surface feature elements of the mining area, and its increase cancause water pollution and vegetation reduction. Therefore, extraction of impervioussurface over mining area is very significant to environmental improvement. Minesurface features are of various types and the environment is complex. How to extractuseful information from mine surface is one of the current difficulties and focuses inacademic research.In this paper, taking Hebi mining area as study area, adopting object-orientedclassification method, using multi-source remote sensing information fusion technologyand combing with GeoEye high-resolution remote sensing images and LiDAR pointcloud data, it carries out land coverage classification and extracts impervious surfaceinformation of mine, which obtains good results. It has some theoretical and practicalvalue, which can commendably provide technical support for the mine’s managementand ecological restoration. The main research contents are as follows:(1)The multi-scale segmentation theory and method are studied, and the imagesegmentation principle, segmentation method, segmentation processes, optimalsegmentation parameters and other issues are analyzed.(2)The study on key technologies of classification instruments selection andclassification rules setup to the object-oriented classification is performed, building aknowledge-based fuzzy classification system and establishing a characteristic selecteddigital quantitative representation model from the shape, spectrum, texture, andrelations between objects.(3)Workflow of impervious surface extraction technology over mining area basedon object-oriented method is built, improving segmentation parameters andclassification rules fit for mining area. Combined with GeoEye high-resolution remote sensing images and LiDAR point cloud data, research of multi-source informationco-processing is conducted.(4)The follow-up treatment and accuracy evaluation of the experimental results arecarried out. The results are evaluated from best classification results, classificationstability and error matrix assessment, which satisfy the classification requirements andgain the rather satisfied impervious surface information over mining area through fieldresearch.
Keywords/Search Tags:object-oriented, impervious surface, high-resolution remote sensing images, LiDAR, multi-scale segmentation
PDF Full Text Request
Related items