The Science Direct website has just published an article detailing the groundbreaking NeuroSYS project’s use of Deep Learning methods to count and estimate the size of shrimp on breeding farms. This significant publication, currently available in open access (pre-proof version), will soon be featured as a full article in the esteemed Journal of Cleaner Production, keeping you at the forefront of the latest advancements in the field.

Recent Publication Reveals AI Breakthroughs in Shrimp Farming

Shrimp production is among the aquaculture industry’s most rapidly expanding sectors. Despite extensive research conducted in recent years, shrimp systems rely on manual sampling to determine stocking densities. This method is neither time-consuming nor expensive, and it threatens the welfare of the shrimp.

Last year, NeuroSYS, in cooperation with the Alfred Wegener Institute from Germany, carried out a successful project (read more in this blog article). It assumed the development of artificial intelligence models for automatically counting shrimp and estimating their size only based on advanced monitoring photo analysis.

AI Success in Commercial Shrimp Farming Systems

The published article describes the course and results of this process. You can learn how NeuroSYS, using eight deep learning-based methods, used artificial intelligence technologies to compare the performance of automated shrimp counting solutions in commercial recirculating aquaculture system (RAS) farming systems.

The work showcases a state-of-the-art object detection method, with the project results demonstrating a remarkable performance in automatic shrimp counting for commercial Recirculating Aquaculture System (RAS) across the entire production range, even in challenging circumstances for object detection. Notably, a high-performing object detection model (YOLOv5m6) was developed, achieving an overall error rate of just 5.97%, reassuring you of the reliability and precision of the technology.

Diagram of the research and technological process of research on the use of AI for automatic counting of shrimps described in the journal of cleaner production
Diagram of the research and technological process of research on the use of AI for automatic counting of shrimps described in the upcoming paper in the “Journal of Cleaner Production”.

About the “Jornal of Cleaner Production”

The Journal of Cleaner Production is a global, multidisciplinary publication dedicated to research and practice in the fields of Cleaner Production, Environmental, and Sustainability. 

It provides a forum for examining and debating both theoretical and practical aspects of cleaner production, covering environmental and sustainability challenges faced by businesses, governments, educational institutions, regions, and societies.

The concept of ‘Cleaner Production’ focuses on minimizing waste generation and enhancing efficiency in utilizing energy, water, resources, and human capital.

Funding Sources

This project, detailed in the upcoming publication in the “Journal of Cleaner Production,” received financial support from the Federal Ministry of Food and Agriculture (BMEL) following a resolution by the German Bundestag. It was sponsored by the Federal Office for Agriculture and Food (BLE) under the innovation promotion program, with the funding code 281C213A19.

Detailed Case Study: Free E-Book Download

For everyone who is interested in aquaculture and wants to learn more about our solution described in the above-mentioned publication. Download the free e-book now and discover all the details about the database, course and results of the NeuroSYS shrimp project. Also explore our solutions and services in the field of AI in aquaculture.

Ai aquaculture cs ebook
From ocean depths to AI heights
Computer Vision in automatic shrimp counting & length estimation