Framework for video-based monitoring of forest insect defoliation and discoloration

Abstract

© The Authors. Pest damage is a general problem that disturbs the growth of forests, influencing carbon sequestration and causing economic losses. In the past decades, many studies have been conducted to monitor and detect forest insect damage using satellite remote sensing technology. Satellite remote sensing has a satellite or aerial vision allowing the monitoring of extensive forest areas, but it usually requires constant time periods and is prone to cloud interference. To enable more efficient and effective monitoring of forest pest damage, a video-based monitoring framework is presented. This framework comprises three key parts: (1) video positioning of forest insect damage based on digital elevation model (DEM) and the parameters obtained from the pan-tilt-zoom camera, (2) integration of two-dimensional/three-dimensional geographic information system and video surveillance to provide more intuitionistic monitoring and assistance for positioning, (3) on-site verification conducted by ground surveys and guided through global positioning system (GPS) integrated in the embedded devices. The experiment was carried out over two forest areas to validate the proposed method. Results showed that the framework bears a sound positioning accuracy and high detection ratio, which could be effectively used in detecting and monitoring forest insect defoliation and discoloration.