Global Risk Assessment Services (GRAS), in collaboration with TH Köln have joined forces to launch the innovative “AI4SusCo” project which is also supported by the Federal Ministry for Economic Affairs and Climate Action (BMWK). The project aims to develop algorithms using ML and AI to improve the classification of forest and coffee plantations and thereby support the compliance of companies with the European Union’s new sustainability regulations.

The primary goal of AI4SusCo is to create a market-ready system capable of globally evaluating deforestation caused by coffee cultivation using EUDR standards. By leveraging state-of-the-art remote sensing technologies and advanced deep learning algorithms, the AI4SusCo project will offer precise mapping and monitoring of coffee plantations and surrounding forests. This innovation will provide coffee importers with reliable automated assessments to meet EUDR compliance, significantly reducing the risk of importing deforestation-linked coffee.

AI4SusCo will employ Convolutional Neural Networks (CNNs) and spatiotemporal methods, to enhance the accuracy of detecting coffee plantations from satellite images. These advanced methods are integrated with various data sources, such as vegetation indices, hydrological, and meteorological data, to refine the mapping and detection processes.

Given the diversity in coffee production, the project will cover the different cultivation techniques, including shaded-coffee and agroforestry systems which are usually not accurately classified using traditional methods. This will help smallholder farmers located in tropical regions who produce coffee in shaded areas to comply with the new regulations.

Also, the AI4SusCo will be customized to meet local requirements, including language options and compliance with regional regulations. This ensures its relevance across diverse geographical regions, enhancing its global applicability concerning deforestation criteria. This also opens potential future services such as certification of deforestation-free cocoa, risk analyses for deforestation, and other products affected by the regulation.

By working together, GRAS and 4C Services can leverage their expertise and capabilities to bring this solution to the market. The partnership with TH Köln enhances expertise in developing complex algorithms and innovative work processes aiding the project’s research capabilities, particularly in the fields of remote sensing, deep learning, and computer vision. This collaboration ensures that the AI4SusCo tool will be widely adopted, offering precise solutions for analyzing cultivation areas concerning deforestation criteria.

Stay tuned for more updates and information about AI4SusCo.