X-band wave radar is a marine environment detection technology with broad application prospects.At present,the extraction of wave parameters using X-band wave radar is relatively mature,but it is not able to reconstruct and invert the wave field.In this paper,a deep learning method based on convolutional autoencoder is proposed to invert X-band wave radar images and obtain two-dimensional wave fields.The error of this method in inverting of JONSWAP spectral wave field with 5.5 m ≤ Hs ≤8.5 m,3.5√Hs ≤Tp ≤4.8√Hs(Hs is significant wave height and Tp is spectral peak period)using this method is within 7％,and this method has good recovery ability for radar images containing noise interference or partial occlusion.This method demonstrates the important significance of inverting two-dimensional wave fields from X-band wave radar images.
关键词
X波段波浪雷达/波浪场反演/卷积/自编码器/深度学习
Key words
X-band wave radar/Wave field inversion/Convolution/Autoencoder/Deep learning
维普
万方数据