| With the rapid development of the national economy, the demand for mineral resources has been increasing in recent years. Mineral resources of high strength exploitation result in destruction and the pollution of mining area land, vegetation, water, and air. With the rapid development of remote sensing technology, timeliness of the mine environment monitoring has been more and more important in ecological civilization construction. Previous studies which mainly used the remote sensing image visual interpretation to obtain mine environment condition are time consuming and laborious, the research of local area fails to take account into activities of mine,map spot of mine and topographic height difference. To solve the above problems,in this paper, in the case of Yunnan Wuding ilmenite mine, based on WorldView-2 and Pleiades image,take account into different mine texture features,topographic height difference and slope to recognize and establish the object oriented technologies which were to used in constructing the classification rules, In 2011,2012,2013,extraction of mine environment information were verified by field identification of study area. By the way, in this paper, author explored to establish the object oriented classification technology of the mining environment in the plateau mountains which could provide a technical reference to the application of the similar areas.In this paper, Achievements obtained are as follows:At first, author summarized the mine object oriented features in the mountain plateau region, which based on the quantitative analysis of the mine object texture, geometry, spectra characteristics from view of objective aspect. The conclusion that, a variety of features combined mines extraction had certain advantages was recongized.Secondly, based on the object attributes, mine environment object spatial semantic features, special features, related features were obtained. The optimal segmentation scale of WorldView-2 method is 125, the optimal segmentation scale of Pleiades is 110. The specific parameters for the study area are listed below,stpoe:shape index is more than 1.5, the brightness value of DEM is more than 2180, the value of slope is more than 3.8, the brightness value of image is less than 712, the value of NDVI is less than 0.25, the value of heterogeneity is less than 0.08 and more than 0.02; transit site: the brightness value of DEM is more than 2123, the value of slope is more than 2.2, the brightness value of image is less than 493 and more than 560, the value of NDVI is less than 0.2, the value of heterogeneity is less than 0.065 and more than 0.031, tailings reservoir:the brightness value of DEM is more than 2213, the value of slope is more than 6.56, the brightness value of image is more than 510 and less than 690, the value of NDVI is less than 0.21, the value of heterogeneity is less than 0.08,the sandard deviation is less than 11; mine building:the brightness value of DEM is more than 2181 and less than 2200, the value of slope is more than 12 and less than 19, the brightness value of image is less than 825 and more than 630, the value of shape index is more than 1.0 and less than 2.1, the value of Density is more than 0.8 and less than 1.8, the ratio of length to width is more than 1.1 and less than 4.5, the value of NDVI is more than 0.04 and less than 0.23.Thirdly, based on the oriented object classification rules which was established above parameters in study area,it is useful to economize manpower that use WorldView-2,Pleiades image which was in 2011,2012,2013 for automatic classification of mine environment. With resort to filed identification, the overall accuracy of the classification results were up to 82.69%,85.89% and 81.67%; and kappa coefficient were 0.8113,0.8467 and 0.8362. It is confirmed that methods and processes for oriented object classification of plateau mountains mine environment are reliable. |
PDF Full Text Request