Seminar on Multispectral Remote sensing images for water depth estimation and landslide detection
International Program Office, Thai Nguyen University of Agriculture and Forestry (TUAF) successfully organised the Seminar on Multispectral Remote sensing images for water depth estimation and landslide detection which represented by Ass. Prof. Hsuan Ren-Center for space and remote sensing research, National Central University . It was held on Wednesday, August 30, 2017 in the central house TUAF .
Dr. Duong Van Thao- Director of International Program Office, Ass. Prof. Dam Xuan Van-Director of CFI, Dr. Tran Thi Pha- Lecturer of Environment faculty-TUAF, Dr. Arinafril- Lecturer of Sriwijaya University and AEP student attended seminar.
Ass. Prof. Hsuan Ren-Center for space and remote sensing research, National Central University
Ass. Prof. Hsuan Ren introduced the advantage of Satellite images. In his study, 2013 island was under water, the next 2 years, they built the airport on this island. The advantage of satellite image is to estimate the water depth and island area without touching.
Taiwan suffered several earthquakes and typhoon which caused landslide each year, infrared image was useful tool to observe landslide and plant cover, . When landslide occurs, coastal line change year by year. Applying satellite image to detect boundary between land and water, water depth ad bottom materials and waterattenuation coeefficient are now becoming more popular.
Due to the economic development and population size, the transportation and natural resource were observed to have better development strategy.
Estimating the water depth can use three methods: SONAR, LIDAR and satellite image (Stereo Pair (image taking from 2 angles), Beer-lambert law, empirical model).
- SONAR: based on speed of sound in water to measure the time difference of transmit and receive sound, it cannot apply in water turbidity. Using the multibeam can only estimate on the track cover.
- LIDAR: up to 50m if water is clear but it is not easy to observe large area.
- Using stereo images taking form different angles and then find the tie points the water (normally it is coral reef). Using the Beer-Lanbert law: sunlight go to water, some will reflect water signal, some will reflect coral reef or sand. Higher error will appear if it is over 30m. Empirical Model: Applying satellite image to estimate water depth and bottom materials.
Formosat 2 can use for natural resource classifications. It needs to have error assessment (difference between Lidar an Neutral network) as well as manual measurement using to verify the images estimation. Using satellite images can observe large areas, especially coastal area, sea grass, live coral, rock, sand, they found that some coral reef did not turn white. It can also observe the scratched coral reef because of boat passing.
Landslide: factors affects the landslide based on rain(direction of speed), surface and underground water; topographic (slope, gradient); geology (fault, rock characteristics); Tectonic (earthquake, volcano); land use (vegetation type, construction). In Taiwan, people do deforestation to grow food… it is the reason landslide can destroy house, road…Applying the satellite to observe land use types.