The Use of UAV Mounted Sensors for Precise Detection of Bark Beetle Infestation

Abstract

The bark beetle (Ips typographus) disturbance represents serious environmental and economic issue and presents a major challenge for forest management. A timely detection of bark beetle infestation is therefore necessary to reduce losses. Besides wood production, a bark beetle outbreak affects the forest ecosystem in many other ways including the water cycle, nutrient cycle, or carbon fixation. On that account, (not just) European temperate coniferous forests may become endangered ecosystems. Our study was performed in the unmanaged zone of the Krkonoše Mountains National Park in the northern part of the Czech Republic where the natural spreading of bark beetle is slow and, therefore, allow us to continuously monitor the infested trees that are, in contrast to managed forests, not being removed. The aim of this work is to evaluate possibilities of unmanned aerial vehicle (UAV)-mounted low-cost RGB and modified near-infrared sensors for detection of different stages of infested trees at the individual level, using a retrospective time series for recognition of still green but already infested trees (so-called green attack). A mosaic was created from the UAV imagery, radiometrically calibrated for surface reflectance, and five vegetation indices were calculated; the reference data about the stage of bark beetle infestation was obtained through a combination of field survey and visual interpretation of an orthomosaic. The differences of vegetation indices between infested and healthy trees over four time points were statistically evaluated and classified using the Maximum Likelihood classifier. Achieved results confirm our assumptions that it is possible to use a low-cost UAV-based sensor for detection of various stages of bark beetle infestation across seasons; with increasing time after infection, distinguishing infested trees from healthy ones grows easier. The best performance was achieved by the Greenness Index with overall accuracy of 78%–96% across the time periods. The performance of the indices based on near-infrared band was lower.

Investigating the Potential Use of RADARSAT-2 and UAS imagery for Monitoring the Restoration of Peatlands

“…We also did not use colour infrared (CIR) imagery. Other studies have successfully used CIR UAS imagery to map and monitor vegetation [80][81][82][83] (because the differences in reflectance between plant species is most easily seen in the near-infrared portion of the electromagnetic spectrum [84]. The authors of [55,81] used CIR in combination with object-based classification to monitor peat bogs.…”

The perspective of unmanned aerial systems in forest management: Do we really need such details?

“…Moreover, the health status of the trees (or, more precisely, early stress detection) provides further key information supporting the economic (Galidaki et al, 2017;Lin et al, 2018) or early detection of pest infestation (Näsi et al, 2018;Lin et al, 2019). On the other hand, a casual digital camera may also be used for reconstructing the forest canopy at very high spatial resolution using Structure from Motion and Multi View Stereo photo-reconstruction techniques (Iglhaut et al, 2019) and GIS software can be employed to derive the information about the number and locations of stands, approximate heights and crown delimitation (Surový et al, 2018); combined with image analysis, classification of tree species or tree stress can be acquired (Fassnacht et al, 2016;Klouček et al, 2019).…”

A Comparison of Multi-Temporal RGB and Multispectral UAS Imagery for Tree Species Classification in Heterogeneous New Hampshire Forests

“…Detailed maps of forest composition are necessary for effective and efficient forest management [1,2]. Maps depicting species-level composition serve a number of applications, such as monitoring biodiversity [3,4], forest health assessments [5,6], conducting precision forestry [7,8] or as inputs for species-specific allometric models [9]. Remotely sensed imagery has been used to decades as a quick and efficient means to produce continuous, large-area maps of forest types [10,11].…”