As of Wednesday (Dec. 10), according to the National Disaster Management Agency (BNPB), the death toll from flash floods and landslides in Aceh, North Sumatra, and West Sumatra has reached 969 people, with 252 others reported missing. The slow pace of search operations underscores the urgent need for breakthroughs in rescue technology.
Responding to this situation, Dr. Andi Dharmawan, an artificial intelligence technology expert at Universitas Gadjah Mada (UGM), said that disaster search-and-rescue efforts could leverage drone-based, computer vision, and smart device technologies.
According to him, such technologies have proven effective in disaster response operations across many developed countries.
Dr. Dharmawan explained that the use of drones for victim search is not new globally. He cited, for example, the deployment of drones during Hurricane Harvey in the United States, where aerial footage was transmitted in real time to speed up evacuations, as well as an incident in Australia in which drones rescued two teenagers swept out to sea by automatically dropping flotation devices.
“Japan and Switzerland have also developed drones equipped with thermal cameras and AI to detect human presence among debris. In Indonesia, drones are being used, but they are not yet integrated with AI. The remaining challenges lie in integration, adoption, and downstream implementation,” he said on Thursday (Dec. 11).
Despite the promise of these technologies, Dr. Dharmawan emphasized that Indonesia’s complex geography poses the greatest challenge. Diverse disaster terrains, ranging from widespread flooding and dense forests to areas without signal coverage, affect drone stability and the accuracy of AI-based detection.
“AI can detect humans in images, but field conditions are not always ideal. Murky water, heavy debris, low lighting, or partially covered victims all pose challenges. The models need to be extremely robust for real-world conditions. In addition, integrating information from drones and AI into on-site SAR teams requires a fast and well-organized system so detection results can be acted upon immediately,” he explained.
To ensure effective implementation, Dr. Dharmawan suggested that technological development should begin with simple applications that deliver immediate benefits.
Stable drones with high-quality real-time video, he said, are already highly useful for search operations without waiting for advanced technologies to be deployed all at once. These can then be supplemented with lightweight AI features to flag areas suspected of containing victims.
“This is not about replacing humans, but about speeding up the process of reviewing video footage. If this stage succeeds, multi-drone systems capable of automatically scanning large areas can begin to be tested on a larger scale,” he said.
Dr. Dharmawan stressed that field trials are the key to successful integration of victim search technologies in Indonesia. Even the most advanced technologies require extensive adaptation if they have not been tested in Indonesia’s rapidly changing weather conditions and unpredictable environments.
He noted that research on UAVs, computer vision, and AI at UGM is already highly advanced but still requires support for downstream implementation and collaboration with disaster management agencies to be truly deployed in official operations.
The expert expressed hope that innovation in victim search would not stop at the conceptual stage. At present, such technologies remain largely at the research level.
Ideally, stronger support is needed so existing research can move toward downstream application and be used in real SAR operations. He also hopes for routine field trials to ensure continuous technological development.
“It doesn’t have to be large-scale right away. What matters is steady progress until it can genuinely help save lives,” Dr. Dharmawan concluded.
Author: Kezia Dwina Nathania
Editor: Gusti Grehenson
Post-editor: Salma
Photograph: Antara