Improved Mangrove Crown Measurement from Airborne Lidar Data using Marker-controlled Watershed Algorithm-filtered Hamraz Technique

Mangrove forest is one of the most essential types of forest in the Philippines. Although they cover the country for about 256185 hectares, they play a vital role in the ecosystem. However, mangroves have faced widespread of destruction throughout the past century. With more than 90% of the worlds, mangroves are located in developing countries like the Philippines and with rapid development and urbanization, these forests are threatened to be destroyed at an alarming rate. With this, there is a critical need to continuously monitor and assess them for dynamic conservation, restoration and management purposes. Moreover, forest inventory, management, and monitoring are very challenging, laborious and expensive to conduct, that is why the emergent of remote sensing application in forest inventory has increasingly become popular. Among the methods in remote sensing, airborne LiDAR has proven effective and reliable. Marker-Controlled Watershed Algorithm, on the other hand, was known as the most common type of technique utilized to conduct the processing of lidar data in terms of crown delineation, tree segmentation, and extraction. Although the marker-controlled watershed algorithm performs well in the forest with open canopy cover like coniferous, they tend to become less reliable in mangrove forest with complex structure and mixed-vegetations. In this study, an enhanced marker-controlled watershed algorithm (MCWS-HTSR) was proposed. The algorithm will use the marker-controlled watershed algorithm to initially extract the individual tree then apply modified Hamraz's tree segmentation routine to further refine the crown delineation process of the mangrove trees giving a more accurate crown measurement. The algorithm will be validated using the LiDAR data obtained from the mangrove forest in Virac, Catanduanes and the data obtained from the actual field survey.