顾及采动后多源数据融合的矿区大比例尺地形图更新方法研究

For China is a mine disaster-prone country, it shows great significance to provide the latest topographic maps quickly and accurately for relieving disaster, cutting losses, and reducing casualties in mining area. It is the trend and a hot issue to use remote sensing images to update the current large scale topographic maps instead of aerial photographs. However, it is of space separate, ground control points (GCP) rare, and GCP vulnerable to mining disturbance for mine area, especially using remote sensing images purely can not keep the geometric accuracy of the elevation. Therefore, based on high-resolution remote sensing images, mining subsidence prediction model and the original topographic maps，the project is scheming to study such key technologies concerned on the method of large scale topographic map updating as immutable features selection and reliability evaluation of the image control points under mining disturbance conditions, the geometric correction in scarce GCP mining area regarding to the mining disturbance, elevation data updating based on mining subsidence prediction model, multi-source and heterogeneous data integration and fusion methods for mine topographic map updates, to build the update process of large-scale topographic map and create its update mode, quickly obtain the low-cost, high-precision and latest topographic data of the mining area, reveal the rules of spatial distribution of the surface features landforms and spatial temporal evolution of the mining area, and provide safeguard for safe mine exploitation, reasonable mine planning, more anti-disaster preparedness and assistant decision-making as well.

我国是矿山灾害多发国家，如何快速、准确提供现势性强的矿区地形图对及时开展灾后救援、挽回经济损失、降低人员伤亡都具有重要的实际意义。用遥感影像代替航空像片是当前进行大比例尺地形图更新的热点和趋势，但对于矿区，其空间分散、可利用地面控制点（GCP）相对稀少、像控点易受采动影响，同时，纯粹采用遥感影像也无法保证高程的几何精度。基于此，本项目拟以高分辨率遥感影像、开采沉陷预测模型及原有地形图等多源数据为基础，采用GIS等方法，研究开采扰动条件下不变地物特征选取及像控点可靠性评价、矿区无（稀少）GCP条件下顾及采动的影像几何纠正、基于开采沉陷预测模型的高程数据更新、面向矿区地形图更新的多源地图数据集成与融合策略等关键技术问题，以求建立矿区大比例尺地形图更新模式，快速获取低成本、高精度的最新矿区地形数据，揭示矿区地表地物地貌空间分布和时空演变规律，为矿山安全开采、矿区规划及防灾减灾、辅助决策提供保障。

本项目以高分辨率遥感影像、矿区开采沉陷预测模型及原有地形图等多源数据为基础，采用GIS等方法，研究了开采扰动条件下不变地物特征选取及像控点可靠性评价、矿区无（稀少）GCP条件下顾及采动的影像几何纠正、基于开采沉陷预测模型的高程数据更新、面向矿区地形图更新的多源地图数据集成与融合策略等关键技术问题，建立了矿区大比例尺地形图更新模式，为快速获取低成本、高精度的最新矿区地形数据，揭示矿区地表地物地貌空间分布和时空演变规律等提供了依据，同时为矿山安全开采、矿区规划及防灾减灾、辅助决策提供了有效保障。具体说来，通过本项目的实施主要在以下几方面取得了突破：1）提出了一套适合煤矿区的遥感影像纠正时地面控制点选取布设方法，并建立了一种基于严密模型与RFM模型的遥感图形几何纠正新方法，为高精度影像几何纠正奠定了基础。2）针对离散小波变换和离散小波框架变换融合方法的缺陷，提出了一种改进的平移不变离散小波变换（SIDWT）影像融合方法，形成了一套优于常规融合算法的煤矿区大比例尺地形图更新的影像融合方案。3）研发了基于概率积分法预测模型的沉陷区高程地形数据更新系统，实现了开采沉陷区地形图高程的智能化快速更新，解决了遥感影像和摄影测量高程精度低、积水区高程更新困难的难题。4）在系统研究地面控制点选取、数量、布设方案和多源遥感影像纠正、融合、变化检测方法及沉陷区高程更新方法等的基础上，建立了适用于煤矿区大比例尺地形图的最优更新模式，实现了矿区地形图的“非接触”测绘。