Fish stocks are among the most endangered components of marine ecosystems. To minimize threats to marine ecosystems and ensure natural and sustainable resource use, monitoring systems must be placed in oceans and seas. The Underwater Fish Observatory (UFO) and the UFOTriNet are two projects initiated by several researchers from marine biology, engineering and industry in Germany between the years 2014 and 2016 and between 2019 and 2023, respectively. The systems collect abiotic as well as camera and sonar data to count fish stocks. This work proposes a method for robust fish counting using sonar data. Activity segmentation as well as object tracking are important steps to successfully accomplish this task. Background subtraction is often used as pre-processing step for stationary fixed sonars. Our proposed method improves this step using band-pass filtering. For the segmentation step, our method utilizes a simple Gaussian distribution model with positional covariances calculated directly on the intensity image. The tracking step is implemented using a classic Kalman Filter that estimates velocity and position of each object in Cartesian coordinates. The sonar detections in the close range of the observable area are compared to the camera detections for validation. Also automatic parameter optimization is used to maximize correlation with camera detections. Additionally, the proposed method is applied to the Caltech Fish Counting Dataset and compared with a deep learning method based on YOLOv5.

In this work, a low-cost low-power navigation solution for autonomous underwater vehicles is introduced utilizing a Micro-Electro-Mechanical System (MEMS) inertial sensor and an ultra short baseline (USBL) system. The complete signal processing is implemented on a cheap 16-bit fixed-point arithmetic microcontroller. For data fusion and calibration, an error state Kalman filter in square root form is used, which preserves stability in case of rounding errors. To further reduce the influence of rounding errors, a stochastic rounding scheme is applied. The USBL measurements are integrated using tightly coupled data fusion by deriving the observation functions separately for range, elevation, and azimuth angles. The effectiveness of the fixed point implementation with stochastic rounding is demonstrated on a simulation, and the the complete setup is tested in a field test. The results of the field test show an improved accuracy of the tightly coupled data fusion in comparison with loosely coupled data fusion. It is also shown that the applied rounding schemes can bring the fixed-point estimates to a near floating point accuracy.

The multi-layer vector approximate message passing algorithm is applied to the problem of channel estimation in underwater acoustic communication addressing the estimation of time-varying channel characteristics. A pre-trained artificial neural network is used as prior for the channel estimation. The communication receiver is designed as a factor graph with a low-density parity check decoder as factor node. The performance of the proposed communication scheme is evaluated by comparing the bit error rates with classical communication schemes. The bit error rate is significantly decreased on strongly time-varying channels. Different datasets are compared for training the channel estimator.

In this work, we introduce Hydroacoustic and Optical Dataset for Oceanic Research (HODOR), a publicly accessible dataset developed for advanced analysis of marine life. It is one of the first datasets to provide synchronized sonar and optical stereo camera videos, comprising over 400 h of continuous recordings per sensor. The purpose of HODOR is to support the development of machine learning and deep learning algorithms for marine applications, multimodal sensor fusion techniques, biomass quantification, and long-term ecological studies. The data were collected with a stationary underwater fish observatory (UFO) in the Kiel Fjord, an inlet of the Baltic Sea in northern Germany, and includes an 86-day subset of the recordings from 2021, which was selected based on automated detections using the YOLOv5 model. IEEE SOCIETY/COUNCIL Journal of Oceanic Engineering (JOE) DATA TYPE/LOCATION Videos, Abundancies; Kiel Fjord, Baltic Sea DATA DOI/PID https://doi.org/10.1594/PANGAEA.980000