基于面向对象分类法和高分一号影像的露天矿区分类技术研究

Classification of open-pit mine area based on object-oriented technology with GF-1 remote sensing image

为了实现高海拔脆弱生态环境下露天矿区的地物信息提取。利用高分一号卫星影像,对青海省天峻县江仓第五露天矿区进行了面向对象结合分形网络演化多尺度分割方法下的信息提取和分类。在充分利用遥感影像空间信息和地物特征的基础上将研究区域地物分为九类,并将分类结果与典型基于像素分类的最大似然法进行了对比。结果表明:面向对象结合分形网络演化多尺度分割方法针对高分辨率遥感影像分类结果质量优良,可以有效减少混合像元的干扰,总分类精度为88.45%,满足实际生产要求。实现了露天矿区的地物分类。该研究成果可为高海拔脆弱生态环境下的露天矿区管理发展提供技术和数据支持。

The purpose of this study was to achieve the land object information extraction of open mining area under high altitude fragile ecological environment.Using the China GF-1 satellite image,the information extraction and classification of multi-scale segmentation method based on the object-oriented combined fractal network evolution was carried out in No.5 open-pit of Jiangcang,Tianjun,Qinghai.Making full use of the spatial information and features of the image,the study area was divided into nine categories,and the results were compared with the maximum likelihood method based on pixel classification.The results show that the object oriented method has better classification effect for high-resolution remote sensing image,and can effectively reduce the interference of the mixing pixels.The total classification accuracy is 88.45%,meeting the actual production requirements.The classification of land objects in open mining areas is realized.The results of this study can provide technical and data support for the management and development of open mining areas under high altitude fragile ecological environment.