基于DS-InSAR的乌达煤田火区长时序地表形变监测与分析

DS-InSAR-based monitoring and analysis of a long time series of surface deformation in the fire area of the Wuda coal field

煤火燃烧不仅浪费了大量煤炭资源,而且严重破坏了火区生态环境,而传统监测方法存在范围小、频率低、成本高、危险大等问题。为此,研究了一种基于分布式目标合成孔径雷达干涉测量(distributed scatterer interferometric synthetic aperture Radar,DS-InSAR)技术的煤田火区监测方法。该方法通过快速同质点识别算法(fast statistically homogeneous pixels selection,FaSHPS)选取同质点,然后利用特征值分解方法对这些同质点进行相位优化,并根据时间相干性获取最终的分布式目标,最后结合短基线集(small baseline subsets,SBAS)InSAR处理步骤解算时序地表形变。以2017年3月—2019年4月63景Sentinel-1A影像为数据源,利用本文方法获取了乌达煤田时序地表沉降,并与临时相干点合成孔径雷达干涉测量(temporarily coherent point interferometric synthetic aperture Radar,TCP-InSAR)技术监测结果进行了可靠性验证,结果表明:两者间的相关系数为0.84,监测点位密度比TCP-InSAR提高1.24倍;乌达煤田存在严重的地表形变现象,研究区域内最大形变速率为-215 mm/a;煤火区在秋冬季节地表形变变化相对较快,且具有多个形变延伸方向及发育程度不同的沉降中心。

Coal fire not only wastes a lot of coal resources but also severely damages the ecological environment of the fire area.However,conventional monitoring methods suffer disadvantages such as a small scope,low frequency,high cost,and great danger.Therefore,this study developed a monitoring method of coal field fire based on the distributed scatterer interferometric synthetic aperture Radar(DS-InSAR)technology.This method successively selects homogeneous pixels using the fast statistically homogeneous pixels selection(FaSHPS)algorithm,optimizes the phases of these pixels using the eigenvalue decomposition method,obtains the final distributed targets based on the temporal coherence,and calculates the time-series surface deformation by combining the small baseline subsets(SBAS)InSAR technique.Taking 63 scenes of Sentinel-1A images from March 2017 to April 2019 as the data source,this study obtained the time series surface subsidence in the Wuda coal field using this method and then verified the reliability of the results by comparison with the monitoring results obtained using the temporarily coherent point interferometric synthetic aperture Radar(TCP-InSAR)technology.As a result,the correlation coefficient between the two methods was 0.84,but the density of monitoring sites obtained using the method proposed in this study was 1.24 times higher than that of TCP-InSAR.The monitoring results show that the surface of the Wuda coal field deforms severely,with a maximum deformation rate of-215 mm/a,and that the deformation occurs more rapidly during autumn and winter and has multiple extensional directions and multiple subsidence centers at varying degrees.