Objectives In the process of data imaging in the airborne array interferometric synthetic aperture radar (ArrayInSAR), due to the influence of system thermal noise, channel amplitude-phase inconsistency, electromagnetic multipath interference and other factors, there is a lot of noise in the airborne ArrayInSAR point cloud, which seriously affects the three-dimensional reconstruction accuracy based on ArrayInSAR point cloud.

Methods Based on the analysis of the noise characteristics of the airborne ArrayInSAR point cloud, we proposed a hierarchical airborne ArrayInSAR point cloud noise suppression（HPCNS）method. First, according to the distribution of noise points and non-noise points in the ArrayInSAR point cloud, the noise is divided into outlier noise, channel amplitude-phase inconsistency errors noise and multipath interference noise. Then, a hybrid filtering algorithm based on adaptive threshold, a bilateral filtering algorithm based on normal vector and curvature improvement, and a multipath noise removal algorithm based on surface information fitting are integrated to process the three kinds of noise, and the hierarchical noise suppression of the airborne ArrayInSAR point cloud in urban areas is realized.

Results In order to verify the effectiveness of HPCNS method, the denoising effect of HPCNS method is compared with other classical denoising methods. The results show that HPCNS method is superior to other classical denoising methods in the aspects of denoising integrity, accuracy and quality evaluation index.

Conclusions The hierarchical ArrayInSAR point cloud noise suppression method can better remove various types of noise in the ArrayInSAR point cloud data, significantly improve the quality of the airborne ArrayInSAR point cloud, provide high-quality data for three-dimensional reconstruction based on the airborne ArrayInSAR point cloud, and better data support for intelligent mapping.