Acoustic monitoring of underwater sounds to optimize aquaculture systems

Aquaculture is one of the fastest-growing sectors of food production worldwide. It encompasses various systems, such as pond farming, flow-through systems, open-sea net pens and closed recirculation systems. However, these facilities face challenges in terms of sustainability, environmental compatibility, animal welfare and health. Forward-looking technologies such as automation and digitalisation offer decisive solutions. In particular, it is of central importance to continuously monitor the condition of the fish and the technical systems of an aquaculture facility. Modern sensor technology is becoming an increasingly valuable complement to traditional methods, supporting specialist staff. Targeted acoustic monitoring enables deviations to be detected early and appropriate measures to be taken efficiently.

Acoustic monitoring of fish populations in aquaculture systems

Visual monitoring (camera-based) often reaches its limits in aquaculture facilities. Poor lighting conditions and water turbidity often allow only limited observation. Essential information about the health status of the fish, the degree of saturation, and possible stress factors thus remain hidden.

Our technology relies on acoustic monitoring to monitor fish stocks. Underwater sounds are recorded using hydrophones and then analyzed using statistical signal processing and AI models. The aim is to derive automatic recommendations for action from this data – an approach known as Passive Acoustic Monitoring (PAM).

By identifying specific acoustic signals, such as fin beats or water movements for activity levels, as well as sucking and swallowing sounds during fish feeding, the behavior of the fish is modeled. Using a comprehensive reference data set that includes not only acoustic data but also specific, manually recorded characteristics of the fish population (number, size/weight, health status, stocking density, activity level, and tank conditions), an AI model is then trained to assign individual acoustic events to specific fish behavior in real time (Acoustic Event Detection, AED).

Acoustic monitoring of technical systems

In addition to monitoring fish populations, the technology also offers potential for technic equipment monitoring. Hydrophones can also be used for early detection of malfunctions in pumps or ventilation systems in order to continuously monitor the operating status. Flow noises in the water or air stream are automatically compared with previously defined target values.

If the acoustic patterns deviate from the expected signals—for example, due to defective pumps, clogged pipes, or changed operating conditions—the system can indicate possible malfunctions at an early stage. In this way, the reliability and efficiency of the technical infrastructure in aquaculture facilities can be increased.

Innovative technology for sustainable aquaculture systems

The methodology is based on our many years of expertise in audio signal processing and in the detection and classification of acoustic events. We are now applying this experience to the new field of aquaculture.

Fraunhofer IDMT is working closely with Fraunhofer IMTE and companies in the aquaculture industry. Through these collaborations, we combine our expertise in audio analysis with excellent expertise and infrastructure in the fields of aquaculture, fish health, and the optimization of aquaculture systems.

Our acoustic monitoring creates the basis for the future-proof, sustainable, and efficient management of modern aquaculture facilities.