CIDCO: mapping marine species using AI
By Conseil de l’innovation du Québec | Published on September 6, 2024
Problem and objectives
Problem
- The Interdisciplinary Center for Development in Ocean Cartography (CIDCO) is a non-profit organization founded in 2002. Based in Rimouski, it specializes in research, development and training in hydrography and marine sciences.
Objectives
- CIDCO is working to modernize hydrography by integrating advanced technologies such as artificial intelligence (AI) and marine robotics to improve the collection and analysis of ocean data.
Proposed solution
Solution
CIDCO has recently invested in IT infrastructures dedicated exclusively to the development of AI, particularly in terms of big data storage and model calculation. This major investment makes it possible to accelerate data processing and develop more sophisticated algorithms for the analysis of marine data. These infrastructures are a major asset for CIDCO, facilitating the development of large-scale projects involving large quantities of data on an international scale.
Principle
Automated systems, such as underwater autonomous vehicles equipped with cameras and sensors, can collect data with high precision and spatial resolution. This makes it possible to obtain detailed information on habitats, species present, their distribution, and underwater structures.
CIDCO uses deep learning algorithms to identify marine species from underwater images.
For example, during a study in Godbout, Quebec, thousands of underwater images were collected and assembled into an orthomosaic, that is to say a large image where each pixel has a geographic coordinate. An AI model then identified species such as starfish and sea urchins. The detected specimens are then analyzed, measured and geolocated.
CIDCO also uses AI to detect ghost fishing gear and underwater archeology; as well as the characterization of benthic habitats.
Role of AI
The algorithms developed by CIDCO make it possible to quickly process an enormous volume of data that would have been impossible to imagine without the possibilities offered by AI. This opens new horizons in terms of ocean mapping and advantageously positions CIDCO internationally.
Impacts
Main outcomes
Extensive coverage: Automated surveys can cover large areas that would be difficult, if not impossible, to assess by divers or manual methods, especially in conditions that are unsafe or inaccessible to humans.
Reduced costs and risks: Automation reduces the need for costly and risky human dives, especially in deep or dangerous environments. This decreases the risk to diver safety and reduces operational costs associated with diving expeditions.
Temporal tracking: Automated systems can be deployed for extended periods, allowing tracking of environmental and biological changes over time. This is crucial for understanding the impacts of climate change, pollution and other factors on aquatic ecosystems.
Standardized data collection: Automation ensures that data is collected in a consistent and repeatable manner, which is essential for data comparability over time and across different sites.
Accelerated data processing and analysis: Using advanced software to analyze data collected by automated systems can significantly speed up data processing. This allows researchers and managers to quickly obtain valuable insights.
Automatic detection and identification: Thanks to advances in artificial intelligence and machine learning, automated systems can automatically identify and classify marine species from collected images and data, reducing the time required for manual analysis.
Challenges takled
Conditions for success
Context of the solution
Market opportunity
CIDCO’s AI projects demonstrate the effectiveness of integrating artificial intelligence into marine research. By automating marine species inventories, detecting ghost fishing gear and characterizing benthic habitats, CIDCO contributes not only to the safety of researchers and the preservation of marine ecosystems, but also to technological advancement in hydrography.
Building on its achievements and its expertise in artificial intelligence, CIDCO proposes to continue this momentum by developing a new project responding to the CRIBIQ call for AI projects.