Rapid identification of shallow inundation for mosquito disease mitigation using drone-derived multispectral imagery
| Autor: | Tasya Vadya Sarira, Philip Weinstein, Lian Pin Koh, Megan Lewis, Ken Clarke |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Health (social science)
Geographic information system 010504 meteorology & atmospheric sciences multispectral 030231 tropical medicine Geography Planning and Development Multispectral image lcsh:G1-922 Medicine (miscellaneous) Breeding 01 natural sciences Normalized Difference Vegetation Index remote sensing 03 medical and health sciences 0302 clinical medicine Animals Ecosystem Drones mosquitoes 0105 earth and related environmental sciences Remote sensing business.industry Health Policy Water Vegetation Drone Waves and shallow water Culicidae Habitat Remote sensing (archaeology) Geographic Information Systems Environmental science business lcsh:Geography (General) Software habitat identification Environmental Monitoring |
| Zdroj: | Geospatial Health, Vol 15, Iss 1 (2020) |
| ISSN: | 1970-7096 |
| Popis: | Mosquito breeding habitat identification often relies on slow, labour-intensive and expensive ground surveys. With advances in remote sensing and autonomous flight technologies, we endeavoured to accelerate this detection by assessing the effectiveness of a drone multispectral imaging system to determine areas of shallow inundation in an intertidal saltmarsh in South Australia. Through laboratory experiments, we characterised Near-Infrared (NIR) reflectance responses to water depth and vegetation cover, and established a reflectance threshold for mapping water sufficiently deep for potential mosquito breeding. We then applied this threshold to field-acquired drone imagery and used simultaneous in-situ observations to assess its mapping accuracy. A NIR reflectance threshold of 0.2 combined with a vegetation mask derived from Normalised Difference Vegetation Index (NDVI) resulted in a mapping accuracy of 80.3% with a Cohen’s Kappa of 0.5, with confusion between vegetation and shallow water depths (< 10 cm) appearing to be major causes of error. This high degree of mapping accuracy was achieved with affordable drone equipment, and commercially available sensors and Geographic Information Systems (GIS) software, demonstrating the efficiency of such an approach to identify shallow inundation likely to be suitable for mosquito breeding. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|