Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM68950
Volume:
10 |
11 |
Nov 2026
Pages:
1-3
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.

Department of Electronics and Communication Engineering
K.Ramakrishnan College of Technology

Department of Electronics and Communication Engineering
K.Ramakrishnan College of Technology

Department of Electronics and Communication Engineering
K.Ramakrishnan College of Technology

Department of Electronics and Communication Engineering
K.Ramakrishnan College of Technology

Department of Electronics and Communication Engineering
K.Ramakrishnan College of Technology
Improper segregation of plastic waste reduces recycling efficiency, and manual sorting is slow, labour-intensive and error-prone. This paper presents a low-cost, vision-based system that classifies plastic waste items in real time into three categories: Polyvinyl Chloride (PVC), Polyethylene Terephthalate (PET) and Other Plastics. An ESP32-CAM module captures images of waste items and streams the frames over Wi-Fi to a laptop running a FastAPI and OpenCV backend, which extracts, crops, resizes (224 × 224 pixels) and normalizes each frame. A Convolutional Neural Network (CNN) based on a pretrained MobileNetV2 backbone, trained by transfer learning on images captured with the ESP32-CAM itself, predicts the category. Predictions whose confidence is below 0.6 are reported as UNKNOWN to avoid unreliable decisions. A web dashboard displays the detected category and confidence score and maintains separate counters for each class, while an optional servo-driven bin mechanism can physically separate the item. The model is evaluated using validation accuracy and a confusion matrix. The proposed design offers an affordable and scalable approach for automated plastic waste segregation.
KEYWORDS
This paper presented a low-cost system for real-time classification of plastic waste into PVC, PET and Other Plastics using an ESP32-CAM, a FastAPI and OpenCV backend and a MobileNetV2 transfer-learning model. Images are captured over Wi-Fi, preprocessed to 224 × 224 pixels, classified, and displayed on a dashboard with per-class counters, with an UNKNOWN output for predictions below a 0.6 confidence threshold. An optional servo or bin mechanism allows automatic separation. The approach uses inexpensive hardware and a lightweight network, which makes it suitable for small-scale and educational recycling set-ups.
[1] Geyer, R., Jambeck, J.R., Law, K.L.: Production, use, and fate of all plastics ever made. Science Advances 3(7), e1700782 (2017) [2] LeCun, Y., Bengio, Y., Hinton, G.: Deep learning. Nature 521, 436–444 (2015) [3] Yang, M., Thung, G.: Classification of trash for recyclability status. Stanford University CS229 Project Report (2016) [4] Bobulski, J., Kubanek, M.: Deep learning for plastic waste classification system. Applied Computational Intelligence and Soft Computing 2021, 6626948 (2021) [5] Yosinski, J., Clune, J., Bengio, Y., Lipson, H.: How transferable are features in deep neural networks? In: Advances in Neural Information Processing Systems 27 (2014) [6] Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., Chen, L.-C.: MobileNetV2: Inverted residuals and linear bottlenecks. In: Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4510–4520 (2018) [7] Shorten, C., Khoshgoftaar, T.M.: A survey on image data augmentation for deep learning. Journal of Big Data 6, 60 (2019) [8] Kingma, D.P., Ba, J.: Adam: A method for stochastic optimization. In: Proc. International Conference on Learning Representations (2015)
Publication Date
11/13/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
0
Total views since publication
PDF Downloads
2
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...