Key points are not available for this paper at this time.
Purpose: This work aims to present Unmanned Aerial Vehicles (UAVs), known as drones, as an alternative for environmental monitoring, considering the difficulty of identifying events of illegal dumping in remote areas, whether due to government omission or the difficulty of accessing these locations. These devices provide several benefits, including portability and the capacity to take pictures in real-time from remote and difficult-to-reach places at a cheaper cost without endangering the operator's safety. Design/methodology/approach: The quantity of data gathered by drones needs an enormous amount of effort to analyze and interpret properly due to Rio de Janeiro's difficult topography and large geographic area. To overcome this challenge, artificial intelligence should be used to automatically evaluate and georeference the photos and videos gathered during the investigation. The present work aims to demonstrate the feasibility of this technique in the early and effective identification of remote areas with illegal solid waste dumping by combining a variety of methods, including the use of a global positioning system (GPS), high-precision cameras and sensors on an electronic device for remote sensing (drone), and image processing and interpretation using computational resources. Findings: Following extensive remote sensing in the selected areas, with the assistance of a UAV, the obtained photos were batch-processed utilizing Deep Learning algorithms and digital image processing techniques. After a significant amount of training, these algorithms were able to automatically identify and classify photos that showed signs of solid waste disposal from those that did not, showing that this rapid-automated segregation produced a 92% efficacy. Discussion: Rio de Janeiro produces a lot of trash, much of it is improperly disposed of and indiscriminately deposited in urban and surrounding regions, on hills, and in difficult-to-reach locations. This issue is particularly worse in areas where the government is absent and ineffective, causing citizens to dispose of their garbage in dangerous and hard-to-reach locations. This study offers a Deep Learning-based method for automatically identifying underprivileged communities that require government intervention to reduce irregular garbage disposal and all of the associated environmental consequences.
Musci et al. (Thu,) studied this question.