Marine Data Analytics: Machine Learning Algorithms to Optimize Seaweed Growth

Aquaculture farming faces challenges to increase production while maintaining welfare of livestock, efficiently use of resources, and being environmentally sustainable. To help overcome these challenges, remote and real-time monitoring of the environmental and biological conditions of the aquaculture site is highly important. Multiple remote monitoring solutions for investigating the growth of seaweed are available, but no integrated solution that monitors different biotic and abiotic factors exists. A new integrated multi-sensing system would reduce the cost and time required to deploy the system and provide useful information on the dynamic forces affecting the plants and the associated biomass of the harvest. As part of the EU funded IMPAQT project a new multi modal seaweed sensing system was developed incorporating a variety of sensor to investigate Seaweed growth parameters. The growth rate of seaweed is significantly affected by wave parameters and sea conditions. The wave characteristics in an aquaculture farm are normally measured using expensive equipment, which is not affordable for many farmers or researchers, and is not easily relocated from place to place to evaluate wave conditions in a variety of locations. This research focuses on developing an artificial neural network that can estimate wave height using acceleration and angular velocity data recorded by a low cost IMU sensor.