Research on Target Detection Optimization in Intelligent Solid Waste Classification Based on YOLOv8 Algorithm

Driven by globalization and urbanization, municipal solid waste has surged. As the front-end link of resource recovery and environmental protection, garbage classification is particularly urgent in order to improve its automation and intelligence level. Based on the YOLOv8 algorithm, this study proposes a series of optimization strategies such as using GhostBottleneck lightweight module, fusion improvement of CBAM attention mechanism, and using WIoU loss function for target detection tasks in intelligent garbage classification systems, aiming at improving detection accuracy and system robustness. This study considered complex issues such as lighting conditions during garbage classification, the integrity of the garbage itself, the severity of the environment, and the difficulty of identification. It used multiple shared garbage image data sets on the network to test the garbage classification optimization model. The average accuracy rate was significantly improved, reaching 96.12%.