With the acceleration of urbanization and increasingly frequent human activities, water-environment pollution has become a growing concern. Floating surface garbage not only destroys the ecological landscape but also threatens the survival of aquatic organisms and the safety of drinking water. Traditional manual patrol inspection is inefficient and limited in coverage and can hardly meet the demand for large-scale dynamic monitoring. By combining computer vision and natural language processing, this project designs and implements a You Only Look Once (YOLO)-based water-area garbage recognition and question-answering system. Using You Only Look Once version 11 (YOLOv11) and Faster Region-based Convolutional Neural Network (R-CNN) as dual core models, the system supports image, video and real-time camera input. It accurately recognizes 11 categories of water-area garbage such as cans, branches and plastics. A Flask back end and a Bootstrap front end build the interactive interface, integrating Leaflet mean average precision (mAP) visualization and Chart.js statistical charts. These enable garbage-distribution heat-zone selection, temporal trend analysis and model-parameter comparison. An intelligent question-answering module is innovatively introduced, which is based on a knowledge graph and a large language model. It lets users query garbage attributes, disposal suggestions and environmental-protection knowledge in natural language. The system combines real-time performance, accuracy (mAP≥0.75) and ease of use, and can be deployed on cloud servers or edge devices. It provides a low-cost and high-efficiency water-area monitoring solution for environmental-protection agencies, water-affairs institutions and the public.
References
[1] Chen, L., Zhu, J. (2024) Water surface garbage detection based on lightweight YOLOv5. Scientific Reports, 14(1), 6133.
[2] Jiang, P., Ergu, D., Liu, F., Cai, Y., Ma, B. (2022) A review of Yolo algorithm developments. Procedia Computer Science, 199, 1066-1073.
[3] Yi, Z., Yao, D., Li, G., Ai, J., Xie, W. (2022) Detection and localization for lake floating objects based on CA-faster R-CNN. Multimedia Tools and Applications, 81(12), 17263-17281.
[4] Ameen, H., Soilán, M. (2026) A review of large language models in geomatics: integrating multimodal data, addressing challenges, and exploring synergies. Spatial Information Research, 34(2), 16.
[5] Hao, C. Y., Chen, Y. C., Chen, T. T., Lai, T. H., Chou, T. Y., Ning, F. S., Chen, M. H. (2024) Synthetic data-driven real-time detection transformer object detection in raining weather conditions. Applied Sciences, 14(11), 4910.
[6] Yang, Y., Liu, X., Tu, X., Lu, Y., Wang, Y. (2025) Automating the construction of environmental policy knowledge graph with large language models. Sustainability, 17(22), 10282.
[7] Purkayastha, K. D., Mishra, R. K., Shil, A., Pradhan, S. N. (2021) IoT based design of air quality monitoring system web server for android platform. Wireless Personal Communications, 118(4), 2921- 2940.
[8] Edozie, E., Shuaibu, A. N., John, U. K., Sadiq, B. O. (2025) Comprehensive review of recent developments in visual object detection based on deep learning. Artificial Intelligence Review, 58(9), 277.
[9] Mohan, S., Kumar, B., Nejadhashemi, A. P. (2025) Integration of machine learning and remote sensing for water quality monitoring and prediction: a review. Sustainability, 17(3), 998.
[10] Gou, J., Yu, B., Maybank, S. J., Tao, D. (2021) Knowledge distillation: a survey. International Journal of Computer Vision, 129(6), 1789-1819.
Share and Cite
Jin, Z., Liu, Z., Wang, Z., Liu, C. (2026) Intelligent Aquatic Environmental Protection: A Visual-perception- based Garbage Recognition and Question-answering Decision System. Scientific Research Bulletin, 3(5), 19-27. https://doi.org/10.71052/srb2024/QUKR5219