[U-05] Geospatial Applications for Natural Resources, Environment and Agriculture

[U05-04] Integrating Multi-sensor Satellite Images for Land Change Detection

*Chia-Che Chang¹, Kuo-Hsin Tseng¹ (1.National Central University)

Keywords:Change Detection, SPOT, Sentinel-2

Due to the increasing need for an industrial supply chain, many factories in Taiwan tend to build in non-designated zones without appropriate management, which may threaten the usage of land in the neighboring residential and agricultural areas. In this situation, periodical monitoring of land use and land cover changes (LULCC) can assist identifying illegal land exploitation. Satellite images can provide large-scale and timely information, but using traditional change detection methods is prone to misjudge seasonal phenology as a changed area, especially in the agricultural zone with cyclic land cover patterns. The purpose of this research is to apply multi-spectral and multi-source satellite images in land use and land cover change detection by collecting Sentinel-2 and SPOT-6/7 images in northern Taiwan in 2017-2021. The methodology includes the coregistration of different images and analyzing the temporal change of each pixel in blue, green, and near-infrared bands. By fitting and predicting reflectance seasonal curve in these five years, it is possible to calculate the threshold to determine which area has been intentionally changed in the latest image. Our findings indicate that the overall accuracy obtained by using the developed approach is higher than 90% based on the validation with field investigation data, and successful reduction of detection error due to natural and seasonal phenological changes.