An Atmospheric Phase Correction Method Based on Normal Vector Clustering Partition in Complicated Conditions for GB-SAR

Atmospheric phase is the main factor affecting accuracy of ground-based synthetic aperture radar. The atmospheric screen (APS) may be very complicated, due to drastic changes in conditions, and conventional correction methods based on regression models cannot fit correct it effectively. Partition a feasible path improve (APC) for complicated APS, but overfitting problem ignored. In this article, we propose clustering partition method, normal vector which can APS more reasonably, then perform APC results. APC, simulation experiments measurement data, suggests that proposed method achieves higher than model-based avoids severe overfitting, realizing balance between credibility. This article verifies feasibility effectiveness using distribution information guide conduct APC.