Taiwan - Intelligent multimodal AI mosquito trap

Issues and Objectives

Current methods for monitoring mosquito density (such as the Breteau index and ovipositor index) rely on manual, regular inspections of water-filled containers in communities. Furthermore, they cannot immediately identify the species of mosquitoes, requiring patrol and monitoring personnel to identify them on-site based on experience or to bring them back to the laboratory for testing, which is time-consuming and labor-intensive.

Current monitoring methods are insufficient to capture live mosquitoes and to draw blood from mosquitoes to test for virus types, making it impossible to control dengue fever outbreaks in a timely manner.

---

Improvement Measures Taken

Dr. Liao Lun-te and his research team at the Institute of Biomedical Engineering and Nanomedicine, National Health Research Institutes, have developed a smart, multimodal AI mosquito trap. The hardware utilizes infrared sensors, a high-resolution macro optical image acquisition system, and temperature, humidity, and carbon dioxide detectors. The software incorporates artificial intelligence algorithms, leveraging a pre-built mosquito image database (containing 10,800 dynamic and static images of various mosquito species). After capturing mosquitoes, the device quickly identifies them as disease-carrying mosquitoes such as Aedes aegypti or Aedes albopictus. The captured live mosquitoes can also be used for laboratory blood sample collection for gene sequencing analysis to determine if they carry the dengue virus and have previously bitten mosquitoes. This device also connects to the cloud and geographic information systems, combining environmental information such as temperature, humidity, and carbon dioxide concentration with information on disease-carrying mosquitoes. This facilitates research on mosquito habits and ecology, allowing for further analysis and assessment of high-risk dengue environments, ultimately achieving effective dengue prevention and control.

Image: AI-powered mosquito control technology (Image source: Reference URL 2)

---

Results

The intelligent multimodal AI mosquito catcher captures mosquitoes in an average time of 0.07 seconds with an accuracy rate of over 90%. It is currently being tested in Tainan, Kaohsiung and other places.

Glossary

Climate change risk assessment: Using effective assessment tools to understand the medium- and long-term impacts of climate change on various sectors, and to propose corresponding adaptation strategies and actions to reduce the risks brought about by climate change.

Case type:

Structural and physical options: engineering techniques or reinforcement of existing structures, integrated technology development

Social options: learning and exchanging disaster knowledge in the community, and developing and applying early warning information.

Institutional options: Establishing institutions at the economic, policy, and legal levels

---

Reference Websites

1. https://nmbdcrc.nhri.org.tw/2020/07/31/%e3%80%8e%e8%9a%8a%e3%80%8f%e9%a2%a8%e8%89%b2%e3%80%8e%e8%be%a8%e3%80%8f%ef%bc%9a%e4%b8%80%e5%a0%b4%e4%ba%ba%e8%9a%8a%e7%9a%84%e6%88%b0%e7%88%ad/
2. https://www.gvm.com.tw/article/55574
3. https://newtalk.tw/news/view/2019-01-04/189836
4. https://www.gbimonthly.com/2019/01/38419/
5. https://www.setn.com/News.aspx?NewsID=480141
6. https://news.cts.com.tw/cts/life/201901/201901041948026.html

---