SAR ship target detection method based on background and scale perception

Abstract

The invention discloses an SAR ship target detection method based on background and scale perception. The method comprises the following steps: forming a plurality of groups of parallel background feature extraction modules according to a preset dilated convolution expansion rate and a preset convolution kernel size, the input end of any group of background feature extraction module being connected to a first convolutional layer of a pyramid convolutional neural network, the output end of the background feature extraction module being connected in series with the output end of the pyramid convolutional neural network to generate an initial detection model; determining a scale perception loss function of the initial detection model according to a bounding box and an anchoring box of the to-be-identified ship target, and performing iterative training on the initial detection model based on the scale perception loss function and the multi-scale image set to determine a trained detection model; and detecting the ship in the SAR image by using the trained detection model. According to the technical scheme, background information is fully utilized, the false alarm rate of SAR image monitoring is effectively reduced, and the detection rate of small targets in the SAR image is improved.

Description

technical field [0001] The present application relates to the technical field of synthetic aperture radar target detection, in particular, to a background and scale-aware SAR ship target detection method. Background technique [0002] Surface ship target detection has been widely used in marine supervision, fishery management, ship search and rescue, marine military and other fields. Synthetic Aperture Radar (SAR, Synthetic Aperture Radar) has the characteristics of all-day, all-weather and large width, which enables it to image vast sea areas in complex environments, and is an important source of information for ocean monitoring and ocean intelligence extraction. With the increasing number of spaceborne SAR and airborne SAR sensors and the increasing data of seaborne SAR, there is an urgent need for an intelligent algorithm to quickly realize SAR ship target detection. Thanks to the rapid development of high-speed parallel computing hardware (GPU, TPU, DPU, etc.) and deep ...

AI Technical Summary

Problems solved by technology

[0004] 1) When detecting ships near coastlines, ports and islands and reefs, strong scattering objects on land will form strong clutter, resulting in a high false alarm rate

[0005] 2) Small and medium-sized fishing boats account for the largest proportion of the total number of ships in the ocean, and small and medium-sized fishing boats are small in size and occupy fewer pixels in the SAR image, while small-sized targets are prone to loss of information during neural network transmission, resulting in higher leakage. Detection rate

Images