矿区沉降监测的InSAR质量图相位解缠方法

A New Quality Guided Phase Unwrapping Algorithm for Coal Mining Area Subsidence Monitoring

煤矿沉降区InSAR干涉图通常具有条纹密集、噪声大、残差点多等特点,传统相位解缠方法难以完全胜任。质量图相位解缠方法在展开此类干涉图相位时比较稳定。不同的质量图指标大多单方面偏重时间相关或空间相关,往往导致解缠效果不理想。提出了一种融合相位导数变化和相关系数的新质量指标,既考虑像元间的时间相干性又兼顾其空间相似性,较适合于矿山沉降区InSAR相位解缠。一组模拟的高山地形干涉图相位解缠结果表明,相较于相位导数变化和相关系数指标,新的质量指标可提高解缠精度,避免大误差累积。一组地物覆盖复杂、噪声较大的实测矿区沉降干涉图的解缠结果表明,新的质量指标较可靠,可有效避免高质量区域误差累积。

The InSAR interferogram of coal mine subsidence area is usually characterized by dense fringes,large noise and many residuals.The phase unwrapping method based on quality map is relatively stable when unwrapping the phase of this kind of interferogram.However,different quality indexes tend to focus on temporal correlation or spatial correlation unilaterally,which often leads to unsatisfied unwrapping results.In this paper,a new quality index combining phase derivative variation and correlation coefficient is proposed,which not only considers the temporal coherence between pixels,but also takes into account their spatial similarity.It is more suitable for phase unwrapping of InSAR in mine subsidence area.A set of simulated interferometric phase unwrapping results indicates that the new quality index can improve the unwrapping accuracy and avoid large error accumulation compared with phase derivative variation and correlation coefficient index.The unwrapping results of a set of real mine subsidence interferograms with complex ground cover and high noise show that the new quality index is reliable,which can effectively avoid the low unwrapping accuracy caused by the accumulation of high-quality regional errors.