SAR image berthing ship target detection method

Abstract

At present, a mainstream SAR ship detection method still has many unsolved difficult problems, such as land and artificial buildings near a coastline, such as plank roads, houses and control towers, strong scatterers cause great interference to ship detection, and partial areas have scattering characteristics similar to those of ships, so that the detection effect of the existing method on ships near the coastline is poor, and the detection efficiency is high. The false alarm rate is high, and the azimuth and contour of an offshore ship are difficult to accurately detect. According to the SAR image berthing ship target detection method provided by the invention, the position, orientation, contour and feature information of the berthing ship near the coastline can be accurately obtained, and the influence of strong scatterers such as land and artificial buildings near the coastline on ship detection can be reduced.

Description

technical field [0001] The invention belongs to the field of radar target detection, and in particular relates to a method for detecting a ship target in a SAR image. Background technique [0002] "Rapid ship target detection method in large scene remote sensing images based on cascaded convolutional neural network" ("Acta Radar") discloses a cascaded convolutional neural network detection framework, which is pre-screened and fully convolutional by the target Network (P-FCN) and Target Accurate Detection Fully Convolutional Network (D-FCN) are cascaded from two fully convolutional networks. P-FCN is a lightweight image classification network, which is responsible for fast pre-screening of possible ship areas in large scene images. It has fewer layers, simple training, and less redundant candidate frames, which can reduce the calculation of subsequent networks. Burden; D-FCN is an improved U-Net network, by adding the target mask and ship orientation estimation layer to the ...

AI Technical Summary

Problems solved by technology

The current mainstream SAR ship detection method still has many unresolved difficulties, such as land and man-made buildings near the coastline, such as plank roads, houses, towers, etc., strong scatterers cause great interference to ship detection, and some areas have The similar scattering characteristics of ships lead to the poor detection effect of existing methods for ships near the coastline, high false alarm rate, and it is difficult to accurately detect the orientation and outline of ships near the coast

Images