Search

Recently, transformers have demonstrated notable improvements in natural advanced visual tasks. In the field of computer vision, transformer networks are beginning to supplant conventional convolutional neural networks (CNNs) due to their global receptive field and adaptability. Although transformers excel in capturing global features, they lag behind CNNs in handling fine local features, especially when dealing with underwater images containing complex and delicate structures. In order to tackle this challenge, we propose a refined transformer model by improving the feature blocks (dilated transformer block) to more accurately compute attention weights, enhancing the capture of both local and global features. Subsequently, a self-supervised method (a local and global blind-patch network) is embedded in the bottleneck layer, which can aggregate local and global information to enhance detail recovery and improve texture restoration quality. Additionally, we introduce a multi-scale convolutional block attention module (MSCBAM) to connect encoder and decoder features; this module enhances the feature representation of color channels, aiding in the restoration of color information in images. We plan to deploy this deep learning model onto the sensors of underwater robots for real-world underwater image-processing and ocean exploration tasks. Our model is named the refined transformer combined with convolutional block attention module (RT-CBAM). This study compares two traditional methods and six deep learning methods, and our approach achieved the best results in terms of detail processing and color restoration. [ABSTRACT FROM AUTHOR] more...

Aiming to address the multi-objective cooperative optimization problem of ship structure lightweight and crashed-resistance, this study proposed a concave reentrant honeycomb-based negative Poisson's ratio (NPR) sandwich structure to replace the stiffened plate. The feasibility of using the sandwich structure to improve ship crashworthiness is demonstrated through finite element analysis. An intelligent optimization method is established to address the issues of multiple parameters and complex effects associated with composite structures and optimize the crashworthiness of the sandwich structure. The effectiveness of the optimization method and the feasibility of the NPR sandwich structure in ship crashworthiness have been validated, providing a reference for the design of new multi-functional ship structures. [ABSTRACT FROM AUTHOR] more...

In this paper, the novel pendulum tuned mass damper model has been suggested as a potential practical engineering modification to an existing offshore crane setup. A number of potential numerical singularities have been circumvented and substantial numerical model accuracy has been reached. Main application of this study would be extreme value statistics, in case of e.g. extreme weather conditions causing collision between vessel and payload. Next, extreme statistics of random vibrations has been studied in this paper for a suggested pendulum tuned mass damper model under white noise excitation. Restoring force has been modelled as the elastic non-linear, and numerical comparison has been performed with the linearised restoring force model case to see the force non-linearity effect on the vessel payload response statistics. The statistics of non-linear case has been studied by applying the path integration method, based on the Markov property of the coupled dynamic system. [ABSTRACT FROM AUTHOR] more...