Mangrove Species Classification in Qi’ao Island Based on Gaofen-2 Image and UAV LiDAR

Mangrove species classification is of great significance to the study of mangrove community structure and biodiversity. Most researches use foreign high-resolution remote sensing images or UAV images for mangrove species classification. In order to improve classification accuracy, LiDAR and hyper-spectral data are often used to assist classification. In this paper, based on the Gaofen-2 image and the CHM data obtained by the UAV Lidar, the mangroves in Qi'ao Island, Zhuhai are classified among species by using the random forest classification method. The classified species include 5 types of true mangroves, 3 types of semi mangroves, Phragmites australis and non-vegetation. The results show that the use of Gaofen-2 image can only effectively distinguish the Sonneratia apetala, Acrostichum aureum and non-vegetation, the accuracy of distinguishing other mangrove species is not ideal; After Gaofen-2 image fusion of CHM data, the classification accuracy of each mangrove species has been significantly improved, with the overall classification accuracy reaching 91.44%, which verifies the effectiveness of Gaofen-2 image fusion of external data in mangrove species classification research.