Study On High Resolution Remote Sensing Image Recognition Method For Ion-type Rare Earth Mining

Ion rare earth is an important strategic resource in twenty-first Century,but also a scarce resource in our country.Gannan is known as the kingdom of rare earth,and ion-type rare earth resources are widely distributed.Driven by economic interests,illegal mining,excessive digging phenomenon is very serious,and causing a lot of damage to the natural environment,and seriously affecting the local people's lives However,for rare earth mines are mostly distributed in remote mountainous areas,high mountains and dense forests,many mining sites,scattered mining areas,and complex environment,it is lead to regulatory difficulties and high regulatory costs which have become a prominent issue to restricting the development of rare earth industry.It is a problem of solving urgent to study how to monitor rare earth mining conditions by fast,accurate,real-time,efficient.Based on the characteristics of rare earth mining and the characteristics of remote sensing image,this paper constructs a method to quickly identify rare earth mining points by using object-oriented method to provide the basis for dynamic monitoring.In this paper,taking the Xunwu County Heling rare earth ore and Dingnan rare earth ore in Gannan as the study area,according to the characteristics of rare earth mining,using Pleiades images and aerial images as the basic data,studying remote sensing identification methods for rare earth mining.First,combining rare earth ore mining technology with field investigation,remote sensing image interpretation symbol of rare earth mining sites with sedimentation tank as the center are constructed.Secondly,colligating image segmentation method of literature at home and abroad,this paper analyzed the advantages and disadvantages,and verified by the experiment,and finally to the comprehensive weighted mean variance method and the maximum area of this research method as the optimal image segmentation scale selection method.Then,based on the characteristics and spatial distribution of various features of rare earth mining area,study various classification rules,spatial relationships between the features,image recognition method to construct the route high spots.Finally,this method was used to identify two kinds of rare earth ore of high resolution images,the recognition accuracy of the sedimentation tank are respectively 94.17% and 90.03%.In summary,in this paper,according to the proposed method,two kinds of high resolution remote sensing image data have been validated by experiments respectively.Experimental results show that the rare earth ore recognition accuracy which come from two kinds of images are above 90%,and the mining sites are being identified.The recognition accuracy of Pleiades satellite images is higher than aerial image,the rare earth ore mining recognition method is more effectively,can provide the source of information for rare earth mines regulation.