Obtaining Riparian Vegetation Characteristics from UAV Optical Imagery 3D Point Cloud Data

River management is an important activity for both ecological protection and the safety of nearby populations. The acquisition of information on vegetation conditions for river management is a difficult task and is often neglected, although recently, it has been facilitated by unmanned aerial vehicle (UAV) technology. This technological advancement, along with artificial intelligence algorithms, has enabled river management professionals and researchers to identify vegetation in riverine areas. Moreover, the use of UAV photogrammetry allows the observation of characteristics, such as vegetation height. This study aims to identify the vegetation patterns and height variation over the course of one year in a 2 km stretch of the Nanakita River in Miyagi, Japan, using monthly UAV-derived 3D point cloud data and artificial neural networks. The vegetation was successfully located for each of the observed months, achieving an accuracy of 98% for spring and summer and 96% for autumn and winter. The area and average height were calculated for each month, and the results showed a pattern of variation of the vegetation amount, demonstrating that summer has a peak amount, as opposed to the winter period. The applied method was effective for the objectives, proving that UAV imagery is an important tool for river management.