Digitalization and Automation

The visual computing research area uses optical technologies combined with artificial intelligence (AI) for automated image data analyses in aquaculture. This technology supports stock and biodiversity monitoring through efficient counting and species differentiation as well as precise monitoring and analysis of key aspects of fish farming, which are critical for animal welfare and efficient farming. AI-based image analysis enables early detection of diseases, parasites and injuries, monitoring of nutritional status and feed intake as well as analysis of behavioral aspects such as stress and social interactions. These insights help to optimize farm operations, improve animal welfare and sustainably increase productivity. Accurate and timely collection of relevant data enables operators to make well-founded decisions that contribute to fish health and efficient fish farming.

• Monitoring of fish stocks and biodiversity through automated detection and classification of fish species in image data
• Automated assessment of fish health and physiology through optical detection of anomalies and abnormal behavior
• Automated assessment of nutritional status through optical measurement of fish in farming facilities
• Visualization of facility information for more effective fish farming
• Processing and enhancement of (underwater) images for environmental monitoring or quality control in fishery and aquaculture

Sonar is a method for locating and analyzing underwater structures using ultrasonic signals. In analogy to airborne radar technology, sonar is the most efficient technology for performing a variety of underwater measurement tasks. Many underwater applications require a high-resolution visualization of the environment. This can be done efficiently and accurately with new 3D sonar systems.

• Continuous biomass measurement
• Measurement possible even in highly turbid water
• Imaging accuracy in the millimeter range
• Imaging without illumination  No light-induced stress for the animals
• Direct size determination of animals from acoustic methods possible
• Imaging of underwater structures for maintenance and repair

Autonomous surface vehicle technologies can be used to perform automated inspections of aquaculture facilities. Such inspections could include the recording of water parameters as well as the examination of aquaculture infrastructure using multibeam sonar. Autonomous vehicles can also be used for feeding operations. Depending on the aquaculture facility, autonomous surface vehicles can carry other sensors to help record and determine fish health.

• Equipping platforms with autonomous functionality
• Autonomous data collection
• Inspection of aquaculture infrastructure

• Autonomous surface vehicle (ASV) with obstacle avoidance and multisensor setup for environmental sensing

Thanks to miniaturized nanoplasmonic sensor platforms (consisting of an in-house designed and manufactured sensor chip on a film, an LED source and optoelectronic scanning electronics), Aquaculture@Fraunhofer enables trace substance monitoring for applications in the wastewater, chemical, pharmaceutical and aquaculture industries.

• Robust, optical sensor principle in a compact format
• Can be integrated on-site as a separate device or online
• Parallel monitoring of multiple analytes possible
• For example, working range of 2–14 μg/L shown for diclofenac
• Measurement interval of 15 minutes

• Detection of trace substances such as pharmaceutical residues and microorganisms (viruses and bacteria) possible
• Quality assurance (process and wastewater), needs-based preparation
• Process monitoring and optimization through digitization

Suitable sensors help to make fish production in aquacultures efficient and gentle. Regular recording of various states of the fish as well as the aquaculture environment makes it possible to adjust the behavior of supply systems, optimize feed portions as well as monitor health changes and, if necessary, treat them in a very short time.

Depending on the application, the advantages of different types of sensors are used: In addition to imaging principles such as computer vision and sonar, passive acoustic analysis methods in particular play an important role here. In this process, the sounds produced by fish themselves and by the aquaculture facilities are recorded and analyzed using acoustic event detection. The interpretation of the analysis is then used as a basis for facility control or fish treatment actions. Acoustic monitoring has the advantage of being independent of underwater light and visibility conditions and of being cost-effective in terms of sensors and algorithms.

In combination with other monitoring methods, acoustic condition monitoring also contributes to a comprehensive digitization and automation of aquaculture, which is able to conserve resources (e.g., feed and personnel), maintain fish health and produce robust and safe food at the same time.

• Development and system integration of acoustic sensor systems
• Acoustic event detection in facilities above and under water
• Acoustic process and production monitoring
• Acoustic sensor node development, incl. IoT integration and optimization for edge computing
• Consulting for acoustic optimization of buildings and facilities

• Various hydrophone systems, including Ambient ASF-1 MKII and OpenAcoustics HydroMoth
• Various sensor node systems for IoT integration and edge computing optimization, including NVIDIA Jetson
• Various specialized acoustic chambers for airborne sound measurements, including an anechoic chamber

Underwater vehicles can be used to effectively monitor aquacultures. Unlike fixed monitoring systems, the existing vehicles are mobile and can move to any location and take measurements there. This ranges from video recording to taking water samples or recording chemical parameters. It is thus possible to make reliable statements about the overall state that are not based solely on fixed measurements. The modularity of the used vehicles also allows sensors to be adapted to customer requirements. The existing systems can be used in both fresh and salt water. 

• Testing and development of mobile fish monitoring prototypes
• Holistic coverage of a habitat through the use of mobile systems
• Use of robust UW vehicles for data collection
• Testing of new sensors
• Investigation of the behavior of fish towards robotic systems

• Various functional UW vehicles (ROV, AUV)
• Environmental sensors for vehicles
• Camera systems (mono and stereo cameras, ZEISS optics)

Development of models for aquaponic systems. This includes the plants, the fish and the technical components as well as the resource requirements (power, water). These models can be used to determine the optimum configuration of the overall system for planning and design in order to maximum yield, taking into account product quality, fish and plant health, and minimizing resource consumption. In addition, the models are used to monitor the operating state (fish and plant growth, climatic conditions, water quality, resource consumption in the system) and to implement optimum operational management with the help of the process control system. Fish and plant growth is measured with optical systems. 

• Creation of simulation models
• Model-based design of aquaponic systems
• Anomaly detection for the operating state
• Model-based development of the optimum operational management and integration into the process control system