面向西部矿区大梯度形变的时序SAR形变监测方法

Deformation caused by mining in Western area has the characteristics of large gradient and wide area, and causes a serious threat to the life and property of local residents. The existing time series InSAR technology can not detect large gradient deformation, this limits its application in the western mining area. Considering the above problems, the proposal intends to propose a method of deformation monitoring with time series SAR for the large gradient deformation in western mining area, including: (1) a deformation model of time series PS-InSAR technology that is suitable for characteristics of mining subsidence is proposed, solving the problem that the deformation model of traditional time series InSAR does not match the actual surface deformation; (2) a monitoring method of full basin combining PS-InSAR and Offset Tracking technology is proposed to solve the problem of large gradient and whole basin deformation which can not be monitored by traditional time series InSAR technology; (3) based on the results of deformation monitoring, a three-dimensional deformation estimation method based on the single SAR image pair is proposed to realize the deformation monitoring of three-dimensional and whole basin under the condition of large gradient deformation in the western mining area. The implementation of this proposal will help improve the accuracy and applicability of time series InSAR technology to monitor the large gradient deformation in mining area, and promote the application and development of the time series InSAR technology in the western mining area.

西部矿区形变梯度大、面积广，给矿区人民的生命财产安全带来了严重威胁。现有的时序InSAR技术无法检测大梯度形变，限制了其在西部矿区中的应用。针对以上问题，项目拟研究建立一种针对西部矿区大梯度形变特点的时序SAR监测方法，主要包括：（1）构建符合开采沉陷形变特点的时序PS-InSAR形变模型，解决传统时序InSAR形变模型与实际地表形变不符的问题；（2）发展PS-InSAR与Offset Tracking技术相结合的大梯度全盆地形变估计方法，解决传统时序InSAR无法监测大梯度全盆地形变的问题；（3）基于形变监测结果，提出结合开采沉陷理论的单一SAR影像源全盆地三维形变估计方法，实现西部矿区大梯度形变条件下的地表三维形变监测。本项目的实施将有助于提高时序InSAR技术监测矿区大梯度形变的精度和适用性，推动时序InSAR技术在西部矿区工程化的应用发展。

西部矿区形变梯度大、面积广，给矿区人民的生命财产安全带来了严重威胁。现有的时序InSAR技术无法检测大梯度形变，限制了其在西部矿区中的应用。针对以上问题，项目开展了以下研究工作：（1）提出一种适用于开采沉陷的InSAR最大可检测形变梯度函数模型。该模型选取矿区开采深厚比和下沉系数作为模型参数，利用该模型可以判断不同分辨率、不同波长和不同入射角的SAR在开采沉陷中的最大可检测梯度。该模型可以作为评价InSAR技术在矿区地表形变监测中适宜性与否的参考标准；（2）建立了矿区InSAR最大形变梯度动态检测函数模型。以已有矿区的地质采矿资料为基础，结合Knothe时间函数和概率积分模型，利用不同空间分辨率的TerraSAR数据，建立了矿区InSAR最大形变梯度动态检测函数模型。该模型无需事先获取SAR影像，利用矿区现有的地质采矿资料即可判断InSAR在矿区不同开采阶段的可检测区间；（3）提出一种联合基于SAR影像幅度信息的像素偏移量跟踪与概率积分函数模型相结合的矿区大梯度形变监测新方法。该方法利用开采沉陷中地表水平移动与倾斜的比例关系，同时结合雷达方程估算垂直、东西与南北三个方向的地表形变；（4）提出一种用于矿区大梯度形变下的地表时间序列三维形变监测与预测方法。该方法对基于SAR像素偏移量跟踪方法得到的多组视线向形变采用多基线反演方法求解得到时间序列的视线向形变，结合开采沉陷模型，利用视线向形变估算垂直、东西与南北三个方向的时间序列形变，同时结合支持向量回归模型进行时间序列的地表三维形变预测；（5）分别从算法的反演准确性、求参稳定性、抗误差干扰能力、全局搜索能力以及运行效率等方面对比分析了模矢法、遗传算法、文化-粒子群算法、粒子群算法、果蝇算法和蚁群算法等六种智能算法。实验结论和分析结果可为后续在进行模型参数（3）和（4）反演时提供参考。