Three-dimensional Point Cloud Extraction Method of Circular Scanning GBSAR Based on Mask Projection and Two-parameter CFAR

‍ ‍Circular scanning Ground Based Synthetic Aperture Radar （GBSAR） rotates around axis has the important advantage of three-dimensional （3D） imaging. However， due to the strong sidelobe in the image caused by the circular observation geometry， it is difficult to extract target point cloud. At present， there is no research on the extraction method of circular scanning GBSAR 3D point cloud. Although the classical two-parameter constant false alarm rate （CFAR） algorithm can extract the 3D point cloud， it needs to detect the 3D image layer by layer at different heights. Due to the influence of strong sidelobes， the 3D point cloud cannot be accurately extracted. Aiming at the above problems， this paper proposes a circular scanning GBSAR 3D point cloud extraction method based on mask projection and two-parameter CFAR. This method uses mask projection and twice two-parameter CFAR to achieve accurate point cloud extraction under the influence of strong sidelobes. This method first applies the maximum projection to the SAR 3D image to obtain the three-view image， then perform the first two-parameter CFAR on the three-views image to obtain the three-view mask. Then， using the three-view mask projection， takes the intersection with the SAR three-dimensional image to remove the spatial position sidelobe and obtain the potential target area data. Finally， the second two-parameter CFAR is applied to the data of potential target area to realize point cloud extraction. The method is tested by actual data and the results show that the proposed method can extract the target point cloud of 3D SAR image more accurately than the classical two-parameter CFAR layer-by-layer detection algorithm.