Deep Learning for Prawn Farming: Forecasting and Anomaly Detection

Joel Janek Dabrowski,Ashfaqur Rahman,Andrew Hellicar,Mashud Rana,Stuart Arnold

DOI: https://doi.org/10.1007/978-3-031-05981-0_3

2022-05-13

Abstract:We present a decision support system for managing water quality in prawn ponds. The system uses various sources of data and deep learning models in a novel way to provide 24-hour forecasting and anomaly detection of water quality parameters. It provides prawn farmers with tools to proactively avoid a poor growing environment, thereby optimising growth and reducing the risk of losing stock. This is a major shift for farmers who are forced to manage ponds by reactively correcting poor water quality conditions. To our knowledge, we are the first to apply Transformer as an anomaly detection model, and the first to apply anomaly detection in general to this aquaculture problem. Our technical contributions include adapting ForecastNet for multivariate data and adapting Transformer and the Attention model to incorporate weather forecast data into their decoders. We attain an average mean absolute percentage error of 12% for dissolved oxygen forecasts and we demonstrate two anomaly detection case studies. The system is successfully running in its second year of deployment on a commercial prawn farm.

Machine Learning

What problem does this paper attempt to address?

The main problem this paper attempts to address is the optimization of water quality management in shrimp farming through deep learning technology. Specifically, the researchers developed a decision support system that utilizes various data sources and deep learning models to provide 24-hour predictions and anomaly detection for water quality parameters. This system aims to help farmers detect water quality issues in advance, allowing them to take preventive measures to avoid risks such as poor growth environments, reduced yields, and even shrimp mortality due to deteriorating water quality. The key issues mentioned in the paper include: 1. **Water Quality Management Challenges**: The most important water quality parameter in shrimp farming is dissolved oxygen (DO). Low DO levels can lead to shrimp hypoxia, suffocation, and even death. Therefore, effectively managing and predicting DO levels is crucial. 2. **Sensor Maintenance Challenges**: Water quality sensors are prone to biofouling under continuous monitoring conditions, which reduces their accuracy or causes damage. This increases the difficulty and cost of sensor maintenance. 3. **Transition from Passive to Proactive Management**: Traditional water quality management methods are mostly corrective after the fact. The goal of this system is to help farmers take proactive management measures through prediction and anomaly detection, preventing problems before they occur. To achieve these goals, the researchers made the following technical innovations: 1. **First Application of Transformer Model for Anomaly Detection**: This is the first time the Transformer model has been used for anomaly detection in the literature. 2. **Extended ForecastNet Model to Handle Multivariate Data**: The ForecastNet model was improved to handle multivariate data. 3. **Incorporation of Weather Forecast Data into Model Decoder**: A new method was proposed to incorporate weather forecast data into the decoders of Transformer and Attention models to improve prediction accuracy. Overall, this paper introduces advanced deep learning technology to provide a more intelligent and efficient water quality management tool for shrimp farming.

Deep Learning for Prawn Farming: Forecasting and Anomaly Detection

Deep reinforcement learning for forecasting fish survival in open aquaculture ecosystem

Pattern detection and prediction using deep learning for intelligent decision support to identify fish behaviour in aquaculture

Temporal fusion transformer-based prediction in aquaponics

Smart Headset, Computer Vision and Machine Learning for Efficient Prawn Farm Management

Algal bloom forecasting with time-frequency analysis: A hybrid deep learning approach.

Enhanced forecasting of chlorophyll-a concentration in coastal waters through integration of Fourier analysis and Transformer networks

Rapid Real-Time Prediction Techniques for Ammonia and Nitrite in High-Density Shrimp Farming in Recirculating Aquaculture Systems

An Innovative Smart and Sustainable Low-Cost Irrigation System for Anomaly Detection Using Deep Learning

Tiny ML-based reconfigurable IoT platform design for brackish water aquaculture monitoring

Deep-Learning-Based Water Quality Monitoring and Early Warning Methods: A Case Study of Ammonia Nitrogen Prediction in Rivers

TSD: Random feature query design for transformer-based shrimp detector

Deep Learning in Aquaculture: A Review

Developing a Novel Water Quality Prediction Model for a South African Aquaculture Farm

A novel fish individual recognition method for precision farming based on knowledge distillation strategy and the range of the receptive field

N-AquaRAM: A Cost-Efficient Deep Learning Accelerator for Real-Time Aquaponic Monitoring

Underwater Fish Image Recognition Based on Detection Transformer

Deep learning model-transformer based wind power forecasting approach

Hybrid deep learning framework for weather forecast with rainfall prediction using weather bigdata analytics

An Integrated Smart Pond Water Quality Monitoring and Fish Farming Recommendation Aquabot System

Deep learning-based prediction of effluent quality of a constructed wetland