Omdena Chapter Page: Indonesia

 Jakarta, Indonesia Chapter Lead – Muhammad Angga Muttaqien

The background

Climate change is increasing the frequency of forest fires and the damage they cause and threatening people with food and water scarcity, increased flooding, extreme heat, more disease, and economic loss. The World Health Organization (WHO) calls climate change the greatest threat to global health in the 21st century and one of the most talked-about consequence of the crisis is forest fires putting millions of our children at risk from air pollution.

Indonesia’s forest fires have made headlines globally over the past few years. According to the latest report in 2019 from the Indonesian National Disaster Management Agency (Badan Nasional Penanggulangan Bencana or BNPB) there are more than 2,000 hotspots in six provinces mainly in Kalimantan and Sumatera Islands. From January to August 2019, 328,724 Ha of land have been burnt by the fire.

Indonesia Fights Fires

The problem

Presidential Instruction No. 11/2015 on forest and land fire prevention is mainly focusing on preventing forest and land fires from happening, fire-fighting, post-fire management and forest and land restoration. This motivates us, as researchers, to seek novel solutions for early fire detection and management. 

Considering the challenges and issues of current methods, using Unmanned Aerial Vehicles (UAVs) for fire monitoring is gaining more traction in recent years. UAVs offer new features and convenience including fast deployment, high maneuverability, wider and adjustable view points. Also, with the advent of Artificial Intelligence especially image-based modeling and analysis we consider AI-powered UAVs can potentially provide extraordinary results.

We are now progressing on building an AI solution for powering drones in disaster relief scenarios and operations such as wildfires to facilitate early fire detection before a catastrophic event happens.

The project goals

1. Collect more annotated fire images and scope them to forest distribution.

2. Build a dedicated model to detect fires in a forest.

3. Optimize the model so it can run on edge devices (drones, UAVs, etc).

4. Publish an interactive dashboard to display how efficient our AI solution to detect fires.

The learning outcomes

1. Understand the problem and current methods to solve the forest fire issue in Indonesia.

2. Participate in collecting relevant datasets for the issue.

3. Engage in the research process of building an AI solution for solving the issue.

4. Educate the community by establishing a sense of urgency regarding the forest fire issue in Indonesia.

The project timeline