We implement practical community-based AI solutions that improve the health and welfare of the communities we live in and the planet we live on. Together with our partners, we deploy these solutions to improve health and welfare, reduce inequality, fight climate change, and aid conservation efforts that protect endangered species and preserve our oceans and forests.
We are working on several exciting projects that utilizes remote sensing data from low earth orbit satellites and computer vision to aid on the ground conservation efforts. Remote areas in the jungle are difficult to patrol and often put a strain on the limited resources of conservationists. Added to that the overall lack of data on changes in the forest and movement of animals make it difficult to accurately count species and measure conservation efforts. With our technology and satellite partners we are utilizing computer vision models to monitor and detect changes in the forest canopy to map the extent and spread over time of invasive plant species. The goal of these projects is to automatically generate quality metrics and data on changes in the forest canopy, prove the viability of utilizing satellite imagery and machine learning models in a cost-efficient and scalable manner to support conservation efforts.
IMPROVE CONSERVATION MANAGEMENT THROUGH MACHINE LEARNING & REMOTE SENSING
Mu Chan Chai Forest
Aiforgood Asia, an NGO focused on AI and technology for environmental and social governance (ESG) projects, partnered with Crayon, an IT service specialist, to address the challenge of forest degradation in Mu Cang Chai forest, Vietnam. The forest is home to endangered species like the western black gibbon, and degradation is caused by illegal cardamon cultivation.
The project aimed to develop a computer vision-based solution that utilized remote sensing and machine learning to detect cardamon cultivation in satellite imagery. Fauna & Flora, an international conservation charity, provided ground truth data for the project.
The project showed that remote sensing and machine learning could effectively detect cardamon cultivation in satellite imagery under certain conditions. The success of this project could help preserve the forest habitat of critically endangered species like the western black gibbon.
The team faced challenges in obtaining cost-efficient remote sensing data for the model, but they eventually used high-resolution images from WorldView-2. The model achieved a high pixel within-sample accuracy of approximately 96%.
The success of this project demonstrates the potential of AI and remote sensing technologies in supporting conservation efforts and addressing complex environmental challenges. It also highlights the importance of collaboration between businesses, organizations, and NGOs to create sustainable solutions for protecting biodiversity and the environment.
HISTORICAL ANALYSIS UTILIZING SATELLITE IMAGERY AND MACHINE LEARNING
Pu Mat National Park
We are collaborating with Save Vietnam Wildlife to map the Pu Mat National Park, located in Nghệ An Province in Vietnam's North Central Coast region. By utilizing historical satellite imagery and Machine Learning technology we aim to categorize the park, such as vegetation types. This data will then be utilized to track changes in the park's composition over time, allowing for future prediction of potential negative impacts the external and internal environment might have on the remaining wildlife in the park.
CONTACT US TO LEARN MORE
Want to learn more about this research project? Please reach out using the contact form below and we'll be in touch: