Models and tools for intelligent image processing in household waste sorting systems
DOI:
https://doi.org/10.34121/1028-9763-2025-3-4-33-53Keywords:
image processing, object recognition, convolutional neural networkAbstract
One of the urgent tasks of environmental management is the efficient recycling of waste based on its sorting. In an intelligent waste sorting system, the image processing algorithm is a key component that determines its efficiency and accuracy. In open datasets, objects in images are usually large, different from those taken in real sorting line conditions. This leads to the lack of samples with small objects. As a result, the model may have difficulty with detection, manifested in missed recognitions or inaccurate positioning. The article combines materials from open sources and photographs obtained in laboratory conditions to create a specialized dataset for visual waste sorting. It included images with a single object, multiple objects, blurred objects, a laboratory background, and a complex background. All images are divided into four main categories: food waste, recyclable waste, hazardous waste, and other waste. The efficient, lightweight YOLO model has been chosen as the base model due to its fewer parameters and high processing speed to achieve low energy consumption and cost efficiency in the intelligent sorting system. The paper analyzes popular image processing and recognition algorithms using convolutional neural networks and, based on them, proposes an improved composite algorithm for garbage detection and classification. A computational experiment has also been conducted to confirm the effectiveness of the proposed algorithm on real training samples. It is shown that the overall performance of the proposed algorithm exceeds current popular object recognition algorithms, ensuring efficient and accurate garbage recognition.
References
1. Коваленко О.Є., Лі Л. Застосування інтелектуальної обробки зображень у системах управління життєвим середовищем. Математичні машини і системи. 2024. № 1. С. 55–69. DOI: https://doi.org/10.34121/1028-9763-2024-1-55-69.
2. Redmon J., Divvala S., Girshick R. You only look once: Unified, real-time object detection. Proc. of the IEEE conference on computer vision and pattern recognition. Piscataway, NJ: IEEE, 2016. P. 779–788.
3. Redmon J., Farhadi A. YOLO9000: better, faster, stronger. Proceedings of the IEEE conference on computer vision and pattern recognition, Piscataway, NJ: IEEE, 2017. P. 7263–7271.
4. Redmon J., Farhadi A. YOLOv3: An incremental improvement. arXiv:1804.02767, 2018.
5. Bochkovskiy A., Wang C., Liao H. YOLOv4: Optimal Speed and Accuracy of Object Detection. arXiv:2004.10934, 2020.
6. Howard A., Sandler M., Chu G. et al. Searching for MobileNetV3. Proc. of the IEEE/CVF International Conference on Computer Vision (ICCV). Piscataway, NJ: IEEE, 2019. P. 1314–1324.
7. Hu J., Shen L., Sun G. Squeeze-and-excitation networks. Proc. of the IEEE conference on computer vision and pattern recognition. Piscataway. NJ: IEEE, 2018. P. 7132–7141.
8. Woo S., Park J., Lee J. CBAM: convolutional block attention module. Proc. of the European Conference on Computer Vision (ECCV). Berlin: Springer, 2018. P. 3–19.
9. Xu B., Wang N., Chen T. et al. Empirical evaluation of rectified activations in convolutional network. arXiv:1505.00853, 2015.
10. Misra D. Mish: A Self Regularized Non-Monotonic Activation Function. arXiv:1908.08681, 2019.
11. Rezatofighi H., Tsoi N., Gwak J. et al. Generalized intersection over union: A metric and a loss for bounding box regression. Proc. of the IEEE/CVF conference on computer vision and pattern recognition. Piscataway, NJ: IEEE, 2019. P. 658–666.
12. Zheng Z., Wang P., Liu W. et al. Distance-IoU loss: Faster and better learning for bounding box regression. Proc. of the AAAI conference on artificial intelligence. Palo Alto, CA: AAAI, 2020. N 34 (7). P. 12993–13000.
13. Ma S., Xu Y. A loss for efficient and accurate bounding box regression. arXiv: 2307. 07662, 2023.
14. Neubeck A., Van Gool L. Efficient non-maximum suppression. 18th international conf. on pattern recognition. Piscataway, NJ: IEEE, 2006. N (3). P. 850–855.
15. Bodla N., Singh B., Chellappa R. et al. Soft-NMS — improving object detection with one line of code. Proc. of the IEEE international conference on computer vision. Piscataway, NJ: IEEE, 2017. P. 5561–5569.
16. Bewley A., Ge Z., Ott L. et al. Simple online and realtime tracking. 2016 IEEE international conference on image processing (ICIP). Piscataway, NJ: IEEE, 2016. P. 3464–3468.
17. Wojke N., Bewley A., Paulus D. Simple online and realtime tracking with a deep association metric. 2017 IEEE international conference on image processing (ICIP). Piscataway, NJ: IEEE, 2017. P. 3645–3649.
18. Thung G., Yang M. TrashNet: A Dataset for Garbage Detection and Classification. GitHub repository, 2017. URL: https://github.com/garythung/trashnet.
126
18
