We use cookies to ensure our website works properly and to personalise your experience. Cookies policy
Department of computer science and engineering, Paavai Engineering College, Paavai Institutions, Paavai Nagar, NH-7, Pachal, Namakkal-637018, Tamilnadu, India.
Urban cleanliness management has become a major challenge in rapidly growing smart cities due to increasing population density and waste generation. Traditional manual monitoring methods are labor-intensive, time-consuming, and inefficient for real-time sanitation management. This paper proposes an AI-based smart city cleanliness detection system using deep learning and computer vision techniques for automatic waste identification and monitoring. The proposed system utilizes CCTV or street camera images and applies the YOLOv8 object detection algorithm to identify garbage accumulation in urban environments. Image preprocessing techniques are implemented to improve detection accuracy under varying environmental conditions. The detected waste regions are analyzed to generate a cleanliness index for different urban areas. Experimental results demonstrate that the proposed model achieves high detection accuracy with real-time performance, making it suitable for smart city applications. The system reduces manual inspection effort and supports municipal authorities in efficient sanitation management.
Smart cities aim to improve urban living standards through intelligent technologies and automated infrastructure management. One of the major challenges faced by modern cities is maintaining public cleanliness and effective waste management. Overflowing garbage bins, roadside litter, and unmanaged waste negatively affect environmental sustainability, public health, and urban aesthetics.
Traditional cleanliness monitoring systems rely heavily on manual inspection by municipal workers. These methods are inefficient, time-consuming, and unable to provide real-time monitoring. With advancements in artificial intelligence and computer vision, automated urban cleanliness detection systems can significantly improve sanitation management.
Deep learning-based object detection algorithms have demonstrated excellent performance in image recognition and environmental monitoring applications. Among these algorithms, YOLOv8 provides high detection accuracy and real-time processing capability, making it suitable for smart city surveillance systems.
This research proposes a smart city cleanliness detection framework using AI and computer vision techniques to identify waste accumulation from street images captured through CCTV cameras or mobile devices. The system generates cleanliness scores and supports municipal authorities in maintaining cleaner urban environments.
LITERATURE SURVEY
Several researchers have explored AI-based waste monitoring systems for smart city applications. CNN-based approaches achieved moderate classification accuracy but suffered from slow processing speed. IoT-enabled smart bins improved waste collection efficiency but lacked large-area cleanliness monitoring capability. Recent deep learning approaches using object detection algorithms demonstrated improved real-time performance. However, challenges remain in achieving accurate waste detection under varying urban environmental conditions. Therefore, this work proposes a YOLOv8-based smart cleanliness monitoring system for efficient urban sanitation management.
PROBLEM STATEMENT
Urban cleanliness monitoring in many cities is still dependent on manual inspection methods, which are inefficient and unable to provide continuous real-time monitoring. Existing systems fail to identify waste accumulation accurately under varying environmental conditions such as lighting changes, weather conditions, and crowded urban scenes. Therefore, there is a need for an automated AI-based cleanliness detection system capable of real-time garbage detection and urban sanitation monitoring.
Manual monitoring is slow and inefficient.
Lack of real-time cleanliness monitoring.
Difficulty in detecting garbage under varying conditions.
Increased urban waste generation.
High labor and maintenance costs.
Delayed municipal response to waste accumulation.
Environmental and public health concerns.
Traditional urban cleanliness monitoring systems rely heavily on manual inspection methods, which are inefficient, time-consuming, and unable to provide real-time monitoring. Existing systems face challenges in accurately detecting waste accumulation under varying environmental conditions. Hence, an AI-based automated cleanliness detection system is required for efficient smart city sanitation management.
PROPOSED SYSTEM
The proposed system captures urban street images using CCTV cameras or mobile devices. The collected images are preprocessed to improve quality and remove noise. The YOLOv8 deep learning model is trained using labeled waste images for garbage detection. The trained model identifies garbage objects and calculates cleanliness scores based on detected waste density.
Camera Input → Image Preprocessing → YOLOv8 Detection → Garbage Identification → Cleanliness Score Generation → Municipal Alert SystemFeature Extraction.
SYSTEM ARCHITECTURE
EXPERIMENTAL RESULTS AND DISCUSSION
|
Method |
Accuracy |
Precision |
Recall |
F1-Score |
|
Traditional Methods |
85.2% |
83.5% |
82.1% |
82.8% |
|
CNN-Based Detection |
91.3% |
90.2% |
89.5% |
89.8% |
|
Proposed YOLO V8 model |
96.5% |
95.2% |
94.8% |
95.0 |
ADVANTAGES
The proposed AI-based system provides automated real-time cleanliness monitoring with higher accuracy and faster detection speed compared to manual inspection and traditional image processing techniques. It reduces human effort and improves smart city sanitation management efficiency.
CONCLUSION
This paper presented an AI-based smart city cleanliness detection system using deep learning and computer vision techniques. The proposed YOLOv8 model successfully identified garbage accumulation in urban environments with high accuracy and real-time performance. The developed system reduces manual monitoring effort and supports efficient urban sanitation management. The proposed framework can be integrated into smart city infrastructure for automated cleanliness monitoring and improved public hygiene.
FUTURE WORK
Future work includes drone-based cleanliness monitoring, IoT integration, GPS-enabled waste mapping, edge AI deployment, predictive waste analysis, and smart dashboard development for large-scale smart city sanitation management.
REFERENCES
Abinesh M.*, NMK Ramalingam Sakthivelan, Smartcity Cleanliness Detection Using Ai Based Techniquies, Int. J. Sci. R. Tech., 2026, 3 (5), 513-515. https://doi.org/10.5281/zenodo.20198821
10.5281/zenodo.20198821