Underwater salient object detection (USOD) attracts increasing interest for its promising performance in various underwater visual tasks. However, USOD research is still in its early stages due to the lack of large-scale datasets within which salient objects are well-defined and pixel-wise annotated. To address this issue, this paper introduces a new dataset named USOD10K. It consists of 10,255 underwater images, covering 70 categories of salient objects in 12 different underwater scenes. In addition, salient object boundaries and depth maps of all images are provided in this dataset. The USOD10K is the first large-scale dataset in the USOD community, making a significant leap in diversity, complexity, and scalability. Secondly, a simple but strong baseline termed TC-USOD is designed for the USOD10K. The TC-USOD adopts a hybrid architecture based on an encoder-decoder design that leverages transformer and convolution as the basic computational building block of the encoder and decoder, respectively. Thirdly, we make a comprehensive summarization of 35 cutting-edge SOD/USOD methods and benchmark them over the existing USOD dataset and the USOD10K. The results show that our TC-USOD obtained superior performance on all datasets tested. Finally, several other use cases of the USOD10K are discussed, and future directions of USOD research are pointed out. This work will promote the development of the USOD research and facilitate further research on underwater visual tasks and visually-guided underwater robots. To pave the road in this research field, all the dataset, code, and benchmark results are publicly available: https://github.com/LinHong-HIT/USOD10K.

Applying swarm intelligence to actual swarm robotic systems is a challenging task especially with adequately consideration of corresponding practical constraints. Under the restrictions of the field-of-view limited relative positioning, local sensing and communication, kinematic limitations as well as anti-collision issues, this paper presents a constrained Particle Swarm Optimization (PSO) based collaborative searching method for robotic swarms. Besides, the proposed method follows the concept of evolution speed and a modified aggregation degree to determine the adaptive weights in the robotic PSO model. The modified aggregation degree is associated with the number of members in one’s field-of-view. Unlike the traditional position update method, the proposed method updates the forward speed and angular velocity of the robot using the non-holonomic model to realize the motion control of each robot. The simulation results show that the proposed method has the potential for the practical implementation of collaborative searching tasks for robotic swarms in different types of environments.

Cable-driven parallel robots (CDPRs) have the characteristics of large workspace and fast response speed, which attracted lots of interests in the past decade. Most of the CDPRs use flexible cables to manipulate the moving platform while the length of cables is changed by the winches fixed on the base. Consequently, interferences among the driving cables, the moving platform, and the environment become inevitable problems, which significantly limit the effective workspace of CDPRs and its performance in complex multiobstacle environment. This article proposes a novel CDPR with movable anchor points (M-CDPR). The movable anchor point of each driving cable can move along the track on the fixed frames. Besides, we establish the kinematics and dynamics of the M-CDPR, and propose the collision detection method and the obstacle avoidance algorithm. Simulation analyzes indicate that the wrench-feasible workspace of the M-CDPR increases 456.06% compared with traditional CDPRs in the same dimension. Moreover, the M-CDPR can realize obstacle avoidance without changing the desired trajectory under different conditions when there exist multiple static obstacles, or both moving and static obstacles. In addition, a prototype is fabricated, consisting of eight driving cables and eight movable anchor points. Experimental results show that the proposed M-CDPR performed well in an environment containing a number of static and/or moving obstacles without any interference.

This paper conducts a numerical investigation on the hydrodynamic performance of a portable autonomous underwater vehicle (AUV). The portable AUV is designed to cruise and perform some tasks autonomously in the underwater world. However, its dynamic performance is strongly affected by hydrodynamic effects. Therefore, it is crucial to investigate the hydrodynamic performance of the portable AUV for its accurate dynamic modeling and control. In this work, based on the designed portable AUV, a comprehensive hydrodynamic performance investigation was conducted by adopting the computational fluid dynamics (CFD) method. Firstly, the mechanical structure of the portable AUV was briefly introduced, and the dynamic model of the AUV, including the hydrodynamic term, was established. Then, the unknown hydrodynamic coefficients in the dynamic model were estimated through the towing experiment and the plane-motion-mechanism (PMM) experiment simulation. In addition, considering that the portable AUV was affected by wave forces when cruising near the water surface, the influence of surface waves on the hydrodynamic performance of the AUV under different wave conditions and submerged depths was analyzed. Finally, the effectiveness of our method was verified by experiments on the standard models, and a physical experiment platform was built in this work to facilitate hydrodynamic performance investigations of some portable small-size AUVs.

Underwater images inevitably suffer from degradation and blur due to the scattering and absorption of light as it propagates through the water, which hinders the development of underwater visual perception. Existing deep underwater image enhancement methods mainly rely on the strong supervision of a large-scale dataset composed of aligned raw/enhanced underwater image pairs for model training. However, aligned image pairs are not available in most underwater scenes. This work aims to address this problem by proposing a novel weakly supervised underwater image enhancement (named WSUIE) method. Firstly, a novel generative adversarial network (GAN)-based architecture is designed to enhance underwater images by unpaired image-to-image transformation from domain X (raw underwater images) to domain Y (arbitrary high-quality images), which alleviates the need for aligned underwater image pairs. Then, a new objective function is formulated by exploring intrinsic depth information of underwater images to increase the depth sensitivity of our method. In addition, a dataset with unaligned image pairs (named UUIE) is provided for the model training. Many qualitative and quantitative evaluations of the WSUIE method are performed on this dataset, and the results show that this method can provide improved visual perception performance while enhancing visual quality of underwater images.