Abstract
As acoustic waves are affected by the channel characteristics, such as scattering and reverberation when propagating in water, sonar images often exhibit speckle noise which will cause visual quality of the image to decrease. Therefore, denoising is a crucial preprocessing technique in sonar image applications. However, speckle noise is mainly caused by the sediment echo signals which are related to the background of seafloor sediment and can be obtained by prior modeling. Although deep learning-based denoising algorithms represent a research hotspot now, they are not suitable for such applications due to the high calculation amount and the large requirement of original images considering that sonar is carried by Autonomous Underwater Vehicles (AUVs) for collecting sonar images and performing calculation. In contrast, dictionary learning-based denoising method is more suitable and easier to be modeled. Compared with deep learning, it can greatly reduce the calculation amount and is more easily integrated into AUV systems. In addition, dictionary learning method based on image sparse representation can effectively achieve image denoising similarly. In order to solve the above problems, we propose a new adaptive dictionary learning method based on multi-resolution characteristics, which combines K-SVD dictionary learning with wavelet transform. Our method has the characteristics of dictionary learning and inherits the features of wavelet analysis as well. Compared with several classical methods, the proposed method is better at speckle noise reduction and edge detail preservation. At the same time, the calculation time is greatly reduced and the efficiency is significantly improved.
Highlights
Acoustic waves have much smaller absorption coefficients underwater than light and electromagnetic waves
In order to solve the above problems, for sonar speckle noise reduction, we propose a new adaptive dictionary learning method based on multi-resolution characteristics, which combine K-singular value decomposition (K-Singular Value Decomposition (SVD)) dictionary learning with wavelet transform
The proposed method has the characteristics of dictionary learning, and inherits the multi-resolution and local characteristics of wavelet analysis
Summary
Acoustic waves have much smaller absorption coefficients underwater than light and electromagnetic waves. Sonar equipment with acoustic waves acting as the carrier plays a very important role in marine monitoring, maritime military operations, and underwater search and rescue. Due to the particularity and complexity of the underwater environment, the echo signals received by sonar are inevitably affected by factors such as channel propagation loss, ocean noise, multipath effects, and reverberation, resulting in the existence of features in sonar images such as low resolution, blurred target edge, and significant speckle noise [1,2,3]. In order to improve the visual effect of sonar images, denoising preprocessing technology has been widely used in feature extraction, target recognition, and image segmentation [4]. According to different processing domains, traditional image denoising algorithms can be generally divided into spatial domain filtering [5] and transform domain filtering methods [6]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
