Extracting Open-Pit Mines From Big Area Based On Google Earth Engine Platform

Open-pit mining has been proven to have a serious impact on the ecological environment and is becoming a focal point in environmental field in various countries as it poses a number of environmental problems that restrict the healthy development of mining areas and surrounding human settlements.In order to effectively regulate and manage open-pit mining and reclamation activities,it is necessary to know the exact extent of the open-pit mining area and how it is changing.Remote sensing is the best way to obtain information on the extent and changes of open-pit mines,however,the current manual visual interpretation of remote sensing images to obtain the extent of open-pit mining areas over large areas is labour-intensive and time-consuming,which greatly hinders the timely understanding of the specific information of open-pit mines.In this study,a "preprocessing-classification-postprocessing" method is proposed for the extraction of large scale open-pit mining areas,and the Landsat and Sentinel series images are used to extract open-pit mining areas in Shanxi Province as the study area.Secondly,the Landtrendr algorithm was used to detect surface disturbances in a wide range of open-pit mining areas to explore the capability of using Landtrendr algorithm to obtain fine scale information on changes in a wide range of open-pit mining areas.The results of this study show that:(1)The "preprocessing-classificationpostprocessing" method proposed in this study can effectively extract open-pit mines over a large area,and the average producer accuracy of the extraction results of openpit mining areas in Shanxi Province from 2000 to 2020 obtained using Landsat series images is 0.794 and 0.832 for each year.The average user accuracy was 0.832 and the average producer accuracy was 0.874,while the user accuracy was 0.923.(2)The Sentinel series images are more suitable for open-pit extraction than the Landsat series images,not only are the extracted open-pit areas more complete,but also able to obtain extraction results year by year,and can obtain specific information on the changes in the open-pit mining area.(3)The Landtrendr algorithm is able to detect mining surface disturbances in a large scale open-pit mining area,thus obtaining fine information on the historical change process in the open-pit mining area,and the fine change information in the large scale open-pit mining area can reflect the overall change trend of open-pit mining activities in the area.The thesis has 41 figures,5 tables,and 69 references.