Underwater vision-based detection plays an increasingly important role in underwater security, ocean exploration and other fields. Due to the absorption and scattering effects of water on light, as well as the movement of the carrier, underwater images generally have problems such as noise pollution, color cast and motion blur, which seriously affect the performance of underwater vision-based detection. To address these problems, this study proposes an end-to-end marine organism detection framework that can jointly optimize the image enhancement and object detection. The framework uses a two-stage detection network with dynamic intersection over union (IoU) threshold as the backbone and adds an underwater image enhancement module (UIEM) composed of denoising, color correction and deblurring sub-modules to greatly improve the framework’s ability to deal with severely degraded underwater images. Meanwhile, a self-built dataset is introduced to pre-train the UIEM, so that the training of the entire framework can be performed end-to-end. The experimental results show that compared with the existing end-to-end models applied to marine organism detection, the detection precision of the proposed framework can improve by at least 6%, and the detection speed has not been significantly reduced, so that it can complete the high-precision real-time detection of marine organisms.

This paper presents an efficient underwater image enhancement method, named ECO-GAN, to address the challenges of color distortion, low contrast, and motion blur in underwater robot photography. The proposed method is built upon a preprocessing framework using a generative adversarial network. ECO-GAN incorporates a convolutional neural network that specifically targets three underwater issues: motion blur, low brightness, and color deviation. To optimize computation and inference speed, an encoder is employed to extract features, whereas different enhancement tasks are handled by dedicated decoders. Moreover, ECO-GAN employs cross-stage fusion modules between the decoders to strengthen the connection and enhance the quality of output images. The model is trained using supervised learning with paired datasets, enabling blind image enhancement without additional physical knowledge or prior information. Experimental results demonstrate that ECO-GAN effectively achieves denoising, deblurring, and color deviation removal simultaneously. Compared with methods relying on individual modules or simple combinations of multiple modules, our proposed method achieves superior underwater image enhancement and offers the flexibility for expansion into multiple underwater image enhancement functions.

Recently, egocentric activity recognition has attracted considerable attention in the pattern recognition and artificial intelligence communities because of its wide applicability in medical care, smart homes, and security monitoring. In this study, we developed and implemented a deep-learning-based hierarchical fusion framework for the recognition of egocentric activities of daily living (ADLs) in a wearable hybrid sensor system comprising motion sensors and cameras. Long short-term memory (LSTM) and a convolutional neural network are used to perform egocentric ADL recognition based on motion sensor data and photo streaming in different layers, respectively. The motion sensor data are used solely for activity classification according to motion state, while the photo stream is used for further specific activity recognition in the motion state groups. Thus, both motion sensor data and photo stream work in their most suitable classification mode to significantly reduce the negative influence of sensor differences on the fusion results. Experimental results show that the proposed method not only is more accurate than the existing direct fusion method (by up to 6%) but also avoids the time-consuming computation of optical flow in the existing method, which makes the proposed algorithm less complex and more suitable for practical application.

Vehicle detection in aerial images is a challenging task. The complexity of the background information and the redundancy of the detection area are the main obstacles that limit the successful operation of vehicle detection based on anchors in very-high-resolution (VHR) remote sensing images. In this paper, an anchor-free target detection method is proposed to solve the problems above. First, a multi-attention feature pyramid network (MA-FPN) was designed to address the influence of noise and background information on vehicle target detection by fusing attention information in the feature pyramid network (FPN) structure. Second, a more precise foveal area (MPFA) is proposed to provide better ground truth for the anchor-free method by determining a more accurate positive sample selection area. The proposed anchor-free model with MA-FPN and MPFA can predict vehicles accurately and quickly in VHR remote sensing images through direct regression and predict the pixels in the feature map. A detailed evaluation based on remote sensing image (RSI) and vehicle detection in aerial imagery (VEDAI) data sets for vehicle detection shows that our detection method performs well, the network is simple, and the detection is fast.

As important observational platforms for the Smart Ocean concept, autonomous underwater vehicles (AUVs) that perform long-term observation in fleets are beneficial because they provide large-scale sampling data with a sufficient spatiotemporal resolution. Therefore, a large number of low-cost micro AUVs with docking capability for power recharge and data transmission are essential. This study designed a low-cost electromagnetic docking guidance (EMDG) system for micro AUVs. The EMDG system is composed of a transmitter coil located on the dock and a three-axial search coil magnetometer acting as a receiver. The search coil magnetometer was optimized for small sizes while maintaining sufficient sensitivity. The signal conditioning and processing subsystem was designed to calculate the deflection angle (β) for docking guidance. Underwater docking tests showed that the system can detect the electromagnetic signal and successfully guide AUV docking. The AUV can still perform docking in extreme positions, which cannot be realized through normal optical or acoustic guidance. This study is the first to focus on the EM guidance system for low-cost micro AUVs. The search coil sensor in the AUV is inexpensive and compact so that the system can be equipped on a wide range of AUVs.

Subglacial lake exploration is of great interest to the science community. RECoverable Autonomous Sonde (RECAS) provides an exploration tool to measure and sample subglacial lake environments while the subglacial lake remains isolated from the glacier surface and atmosphere. This paper presents an electronic control system design of 200 m prototype of RECAS. The proposed electronic control system consists of a surface system, a downhole control system, and a power transfer and communication system. The downhole control system is the core element of RECAS, and is responsible for sonde status monitoring, sonde motion control, subglacial water sampling and in situ analysis. A custom RS485 temperature sensor was developed to cater for the limited size and depth requirements of the system. We adopted a humidity-based measurement to monitor for a housing leak. This condition is because standard leak detection monitoring of water conductivity may be inapplicable to pure ice in Antarctica. A water sampler control board was designed to control the samplers and monitor the on/off state. A high-definition camera system with built-in storage and self-heating ability was designed to perform the video recording in the subglacial lake. The proposed electronic control system is proven effective after a series of tests.