Anthropogenic Marine Debris assessment with Unmanned Aerial Vehicle imagery and deep learning: A case study along the beaches of the Republic of Maldives

Unmanned Aerial Vehicles for Debris Survey in Coastal Areas: Long-Term Monitoring Programme to Study Spatial and Temporal Accumulation of the Dynamics of Beached Marine Litter

“…Our pilot test to use UAVs for monitoring spatial and temporal dynamic of BML accumulation in coastal areas that started in April 2019, proved to be a useful procedure. To our knowledge, this was the first case of using the aerial survey methodology through UAVs to monitor the presence of BML on Italian beaches ( Figure 7 of [40]) and the first case of using UAVs to estimate the accumulation rates of BML on beaches in general. The results of our work showed that the dynamics and the equilibrium of the accumulation process depended, in general, on season, but also on the size of the specific BML.…”

“…The longer battery life, the ability to plan automatic flights with easy-to-use ground station software, and their small size are real advantages, and the structure from motion algorithms (SFM) allow accurate digital elevation models (DEM) and ortho-mosaic terrain models over large areas. Today UAVs are increasingly accessible and have widespread applications, such as in environmental monitoring systems for agroforestry, structural geology, archaeology, marine habitats, supervised hazards, and accidents [23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39], and recently also in monitoring ML on the coast [40][41][42][43][44] or that floating in rivers [45]. These studies are not uniform with regard to the data processing procedures, ranging from visual interpretation of images [42] and analysis of the spectral profile of litter [46], to the use of machine learning methods [43,44].…”

Iceland, an Open-Air Museum for Geoheritage and Earth Science Communication Purposes

“…In most of these works, the use of UAVs has been integrated by Structure-from-Motion (SfM) photogrammetry [89][90][91], a powerful technique to augment traditional methods used to gather outcrop data. Therefore, the combination of UAV-digital image collection and SfM photogrammetry has been increasingly applied to geological and environmental research [39,40]. In order to accomplish the goals of the present work, we have used the above techniques both for collecting images of the selected geosites, as well as constructing 3-D DOMs, which may be considered "Virtual Outcrops (VOs)" [92,93].…”

“…For the sake of the present work, we chose multirotor vehicles (Figure 5a,b), as they can be remotely controlled (e.g., Figure 5c), are characterized by very stable hovering, can be easily transported in the field and are less expensive than hybrid and fixed-wing models (e.g., Figure 5d,e). In addition to the above, multirotor models can fly at very low heights, thus, obtaining greater field resolution; most importantly, take-off and landing operations are smoother than for fixed-wing models, and this can prove to be crucial, especially when operating in difficult logistic terrains, such as lava flow outcrops or remote beach areas [39,40]. Considering the above, we selected two different types of multirotor UAVs: the DJI Phantom 4 PRO quadcopter and the smaller DJI Spark.…”

“…With the purpose of showcasing a total of 25 geosites selected in the two areas, we made use of the following tools: (i) field photographs, some of which have been reworked by highlighting the most relevant features [6]; (ii) highly detailed images, captured by unmanned aerial vehicles (UAVs) and (iii) 3-D models of field outcrops, based on individual, UAV-captured pictures, further elaborated by using Structure-from-Motion photogrammetry techniques (SfM). The importance of UAV-based techniques as innovative and helpful tools for Earth Science research has been widely demonstrated in the last decade, as documented by Bonali et al [39] and Fallati et al [40].…”