广西科技大学学报2025，Vol.36Issue(4)：84-96,113,14.DOI:10.16375/j.cnki.cn45-1395/t.2025.04.011
基于深度学习的平方根容积卡尔曼滤波多基线相位解缠算法
A square root cubature Kalman filtering multi-baseline phase unwrapping algorithm based on deep learning
摘要
Abstract
Phase unwrapping is the key procedures for interferometric synthetic aperture radar technology.In the multi-baseline phase unwrapping algorithm,the cluster-analysis-based multi-baseline phase unwrapping algorithm is a promising and efficient algorithm,but its noise robustness needs to be further improved.For this problem,a square root cubature Kalman filtering multi-baseline phase unwrapping algorithm based on deep learning was proposed.First,a multi-channel edge detection network was constructed for extracting edge information from multi-channel interferograms.Second,the multi-channel interferograms were divided into continuous and mutant regions based on the edge information,and a cluster-analysis algorithm was used to elevation reconstruction for the discontinuous terrain,while the continuous terrain was predicted using a square root cubature Kalman filtering.The experimental results show that the method exhibits high elevation reconstruction accuracy and improved noise robustness.关键词
卡尔曼滤波/相位解缠(PU)/干涉合成孔径雷达(InSAR)/多基线/聚类分析/深度学习Key words
Kalman filtering/phase unwrapping(PU)/interferometric synthetic aperture radar(InSAR)/multi-baseline/cluster-analysis/deep learning分类
信息技术与安全科学引用本文复制引用
田冲宵,谢先明..基于深度学习的平方根容积卡尔曼滤波多基线相位解缠算法[J].广西科技大学学报,2025,36(4):84-96,113,14.基金项目
国家自然科学基金项目(62161003) （62161003）
广西自然科学基金项目(2023GXNSFAA026024)资助 （2023GXNSFAA026024）
