Enhanced surface deformation monitoring in open-pit mines via dynamic filtering and multiscale adaptive interpolation of D-InSAR/offset tracking fusion

D-InSAR is sensitive to noise in low-coherence areas, while offset tracking lacks accuracy, making both insufficient for complex mining monitoring. Robust methods for stable monitoring of highly dynamic, large-gradient deformation remain urgently needed. This study proposes a multiscale adaptive interpolation and dynamic filtering method that fuses D-InSAR and offset tracking techniques to enhance deformation monitoring in mining areas. The method uses multiscale Gaussian decomposition to separate global and local deformation features and applies adaptive dynamic filtering, which employs edge detection and adaptively switches between NLMS and LMS filters, based on the relative difference ratio to suppress noise and enhance edge details. At each scale, data are classified by local variance into high- and low-variance regions, with kriging and linear interpolation applied respectively to balance detail preservation and smooth transitions. Multiscale reconstruction further improves monitoring accuracy and spatial consistency. Experimental results show that in high-gradient mining areas, the proposed method achieves a maximum coherence improvement of 0.22 and a minimum of 0.09, while the phase gradient magnitude increases by up to 0.32, significantly enhancing the expression of deformation details in critical zones. This approach shows strong potential for broader application in complex geological settings and multisource SAR monitoring.