Research On Remote Sensing Recognition Method Of Ground Subsidence Caused By Mined-out Space

Mined-out ground subsidence is an exogenous geological disaster, it destroyed the farmland, damaged the ground buildings, pose a serious threat to industrial and agricultural production and a serious impact on the ecological environment, resulting in a very prominent social and economic contradictions. It is the key for comprehensive management of mine environment and subsidence land reclamation to obtain the subsidence information fast and accurately, the traditional investigate means is time and vigour, so it is difficult to obtain timely and accurate information of subsidence land, and remote sensing images can be record the region ground situation actully, it has obvious advantages to investigate the subsidence land. This paper systematically studied the method of the mined-out ground subsidence with remote sensing images, purpose to be able to obtain the subsidence information timely and accurately, provide basic information for the reclamation and remendiation of mined-out subsidence.Firstly, it analysised the formation mechanism, main types and characteristics of mined-out ground subsidence, and detailed introduced the parameters and technical indicators of commonly used earth resources satellites, summarized the application scope of different data sources in remote sensing recognition of subsidence land, through a number of studies, comparison, summarized. Then summed up the general workflow of remote sensing images pre-processing, the fusion methods were mainly studied. Remote sensing images of Daliuta ground fissure, Xinmi water accumulated subsidence basin, and Huangling collapse were selected reapectively, fused with principal component analysis (PCA), Brovey Transformation, Multiplicative, IHS transformation, High-Pass Filtering(HPF), Gram-Schmidt and Pansharp seven different ways, fusion results were compared by qualitative and quantitative means, through comparison, the effects of Gram-Schmidt and Pansharp are the best. After fusion, different image enhancements were done respectively to images, through comparing test, the Linear Stretching, High-Pass Filtering and Sharpening are more effective methods to enhance. In addition, it summarized the general principles and process of image classification, introduced the common classification methods, and extracted information of water basins, collapses and landslides respectively with image threshold segmentation, the results are satisfactory. It summed up the direct and indirect signs of subsidence pits, ground fissures, landslides and collapses interpretation in the remote sensing image. On this basis, took Xinmi Coal Mine in Henan Province and Daliuta Coal Mine in Shenfu as study areas, subsidence lands in these areas have been recognized with remote sensing images, and the recognization results were evaluated according relevant information, remote sensing recoginition of water basins is simple, but ground fissures only in the larger, low vegetation coverage will be reflected in high-resolution images such as IKONOS and QuickBird.