Year: 2025
Type: article
Abstract: Accelerated growth of urban agglomerations in Kazakhstan causes acute environmental problems associated with pollution of air, water and land environment of cities and suburbs. The number of unauthorized garbage and household waste dumps is growing. Increased generation of municipal solid waste (MSW) is a serious environmental problem, especially in urbanized and industrial regions. Traditional waste monitoring methods are labor intensive and limited in scope. This study proposes a system for automatic MSW detection based on the use of UAV-derived imagery and YOLOv8 deep learning model. The system was tested and achieved high MSW detection accuracy of Precision 84,8 %, Recall 81,5 % and [email protected] (85,1 %), with an IoU of 0,5. Visual analysis confirmed the model's ability to effectively detect waste under different conditions with minimal errors. The research findings can be integrated into environmental monitoring systems for rapid detection of unauthorized waste dumps and optimization of remediation measures.
Source: Hydrometeorology and ecology
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4411825765