Taking a closer look at invasive alien plant research: A review of the current state, opportunities, and future directions for UAVs

Abstract

The development and proliferation of unmanned aerial vehicles (UAV) in recent years presents a new data collection opportunity for invasive alien plant (IAP) research. The flexibility and cost‐efficiency of these craft offers a valuable solution where high‐spatial or high‐temporal resolution remotely sensed data are required. In this paper, we review all published studies using UAV for remote data collection in IAP research. We have systematically identified the taxonomy and habitat characteristics of the system studied, classified the UAV configuration, analytical methods and the limitations of each study. We used this synthesis to identify research gaps, suggest directions for future research, and identify opportunities for practical application of the technology.

Quality Assessment of Photogrammetric Methods—A Workflow for Reproducible UAS Orthomosaics

“…In the future, the general workflow should utilize only open-source software. Currently, Agisoft Metashape is the de facto standard and the most promising software in affordable UAS image processing [3,13]. The development of open-source photogrammetry projects like OpenDroneMap are promising and will be integrated once they are fully operational.…”

“…Unmanned aerial systems (UAS) are widely used in environmental research. Applications encompass the retrieval of crop yield [1] or drought stress [2] in agricultural areas or the mapping of plant species [3][4][5], biomass [6,7] or forest structure [8][9][10] in nature conservation tasks. Today, UAS allow an extensive spatial coverage with high resolution that provides detailed observations on the individual plant level, e.g., for the detection of pest infections in trees [11] or rotten stumps [12].…”

“…The point cloud is the basis for the generation of a surface model and the orthorectification and mosaicing of the individual images based on this surface. Commercial software such as Metashape (Agisoft LLC, St. Petersburg, Russia; formerly known as Photoscan) makes these complex photogrammetric methods accessible to a broad range of users and has been utilized in numerous studies [3,13,24].…”

Detecting plant species in the field with deep learning and drone technology

“…In these studies, plants are detected in digital images, either manually or using automated analysis techniques (Dash, Watt, Paul, Morgenroth, & Hartley, 2019). Unmanned aerial vehicles, commonly referred to as aerial ‘drones’, have provided a more readily available source of high‐resolution imagery and object‐based image analysis (OBIA) has become the dominant analysis technique for detection of targets in this imagery (Blaschke, 2010).…”

“…The predominant workflow for the production of distribution maps, after the collection of drone sourced high‐resolution imagery, currently involves the creation of an orthomosaic and the subsequent application of OBIA with a machine learning classifier for automated classification (Alvarez‐Taboada, Paredes, & Julián‐Pelaz, 2017; Chabot, Dillon, Shemrock, Weissflog, & Sager, 2018; Mafanya et al., 2017; Martin et al., 2018; Wu et al., 2019). However, the intensive processing necessary to produce mosaics acts as a bottleneck, restricting further research (Dash et al., 2019). The end product of such analysis is a map showing the prevalence, distribution and density of the species of interest, which can help support land management decisions and monitoring of populations.…”