一种基于GVF-Snake模型边界探测的相位解缠算法

发布时间：2018-05-13 00:00

本文选题：InSAR + 相位解缠　；参考：《中国矿业大学学报》2017年06期

【摘要】：针对矿区沉降监测中低相干和强噪声下出现相位不连续、条纹间断等导致的InSAR(Interferometric SAR)监测可靠性不高等问题,提出了一种基于GVF-Snake(gradient vector flow snake)模型干涉条纹边界探测的相位解缠方法.该算法利用GVF-Snake模型对沉降漏斗区的干涉条纹进行边缘检测获取干涉条纹的边界,对检测边界内的像元按照相关准则进行解缠绕,通过改进的中值滤波方法消除孤立点.以济宁矿区为实验区,采用日本ALOS PALSAR真实雷达数据进行矿区沉降监测以验证新解缠算法的有效性,利用两组实验数据分别采用基于GVF-Snake模型的边界探测解缠算法与最小费用流等其它5种解缠算法进行对比实验与分析.结果表明:所提算法在干涉条纹质量较好的情况下与其它解缠算法高度一致,以MCF(minimum cost flow)最小费用流解缠方法作为评价标准,新方法解缠精度可以达到±0.005 8rad;在干涉条纹间断、噪声影响严重的情况下,可以实现InSAR矿区沉降漏斗的正确相位解缠,从而保证InSAR矿区大量级沉降监测的可靠性及精度.
[Abstract]:A phase unwrapping method based on GVF-Snake(gradient vector flow snake) model is proposed to solve the problem that the phase discontinuities and fringe discontinuities in mining subsidence monitoring are caused by phase discontinuities and fringe discontinuities. A phase unwrapping method based on GVF-Snake(gradient vector flow snake) model is proposed. The algorithm uses the GVF-Snake model to detect the edge of the interference fringes in the settlement funnel area to obtain the edge of the interference fringes. The pixels in the detection boundaries are unwound according to the relevant criteria and the outliers are eliminated by the improved median filtering method. In order to verify the effectiveness of the new unwrapping algorithm, the ALOS PALSAR real radar data of Japan are used to monitor the subsidence of the mining area in Jining mining area. Two sets of experimental data are used to compare and analyze the boundary detection unwrapping algorithm based on GVF-Snake model and the other five unwrapping algorithms such as minimum cost flow. The results show that the proposed algorithm is highly consistent with other unwrapping algorithms when the quality of interference fringes is good. The accuracy of the new method can reach 卤0.005 radwith the minimum cost flow unwrapping method of MCF(minimum cost flow as the evaluation standard, and when the interference fringes are interrupted, the unwrapping accuracy of the new method can reach 卤0.005 rad. The correct phase unwrapping of the subsidence funnel in InSAR mining area can be realized when the noise is seriously affected, thus ensuring the reliability and accuracy of the monitoring of a large number of levels of settlement in the InSAR mining area.
【作者单位】：山东科技大学测绘科学与工程学院海岛(礁)测绘技术国家测绘地理信息局重点实验室山东省高校海洋测绘重点实验室;
【基金】：山东省自然科学基金项目(ZR2017MD020) 国家自然科学基金项目(41404003) 海岛(礁)测绘技术国家测绘地理信息局重点实验室开放基金(2015B05) 山东科技大学研究生创新基金项目(SDKDYC170313)
【分类号】：TD325.4;TN957.52

【相似文献】