一种InSAR干涉图大气相位融合及校正方法
By combining quadratic polynomials and BP neural networks to simulate atmospheric phase and performing weighted fusion at deformation boundaries, the accuracy problem of atmospheric phase removal in InSAR technology is solved, thus improving the accuracy of deformation monitoring.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- JIANGSU WATER CONSERVANCY SCI RES INST
- Filing Date
- 2023-09-11
- Publication Date
- 2026-07-17
AI Technical Summary
When monitoring land subsidence, existing InSAR technology often uses atmospheric phase removal methods that cannot accurately remove complex atmospheric phases, and polynomial fitting methods have low simulation accuracy. Furthermore, BP neural networks are prone to misfitting deformation phases in deformation regions.
The atmospheric phase in the non-deformed region is simulated using a quadratic polynomial method, while the atmospheric phase in the deformed region is simulated using a backpropagation neural network. The results of the two methods are then fused by weighted averaging at the deformation boundary to form the final atmospheric phase map.
It improves the accuracy of atmospheric phase correction, solves the problems of low accuracy of polynomial fitting and overfitting of BP neural network in deformation region, and enhances the accuracy of InSAR deformation monitoring.
Smart Images
Figure CN117111005B_ABST