Deep learning SAR image ship identification method based on self-supervision condition

Abstract

The invention relates to a deep learning SAR (Synthetic Aperture Radar) image ship identification method based on a self-supervision condition. The method comprises the following steps: firstly, preprocessing SAR data, acquiring an image pixel threshold value by utilizing accumulative inverse exponential probability distribution, carrying out rapid segmentation by utilizing the threshold value to obtain a binary image, then carrying out eight-neighborhood communication processing on the binary image, acquiring geometric information of a candidate target, constructing an SAR ship slice data set according to the geometric information of the candidate target, finally, establishing a CNN model, and training and tuning the CNN model so as to be used for self-supervision identification on the ship target. According to the CNN model based on the self-supervision thought, only a small number of training samples need to be labeled in the recognition process, the sample labeling time is greatly shortened, and the ship detection efficiency is improved; and the backbone model adopts a lightweight model Shufflenet network, model parameters are few, high training precision can be obtained with short training time, the convergence speed is high, and the precision is high.

Description

technical field [0001] The invention belongs to the field of SAR image ship target detection, in particular to a deep learning SAR image ship recognition method based on self-supervision conditions. Background technique [0002] In recent years, with the unprecedented prosperity of ocean trade and the rapid development of ocean transportation, the pressure on world ocean security is increasing, and the possibility of ocean accidents and related environmental damage is also increasing. Traffic management, environmental protection, military security are of great significance. [0003] Remote sensing technology is to obtain the surface information of the earth space by transmitting electromagnetic waves to the surface through space-borne and airborne platforms. In recent years, with the rapid development of remote sensing technology, all-weather, high-resolution spaceborne synthetic aperture radar (Synthetic Aperture Radar, SAR) remote sensing has received widespread attention...

AI Technical Summary

Problems solved by technology

However, there are many defects in the CFAR ship detection algorithm: First, the CFAR algorithm needs to set the target window, protection window and background window, which greatly increases the amount of calculation; Second, with the improvement of the SAR image resolution, the clutter background of the SAR image , Accurate modeling, and detection accuracy are difficult to meet the requirements; third, the gray value of the SAR target fluctuates within a certain range, resulting in incomplete or even broken detection results; fourth, since CFAR is based on the intensity value of a single pixel Judging whether it is the target pixel, and the inherent speckle noise of SAR images is likely to cause a large number of false alarms in the detection results

SAR data is different from optical data. The high-precision labeling of samples requires rich professional knowledge, so it requires a lot of material and brain power, and it is impossible to quickly identify radar targets.

Images