Computer Vision-based Anomaly Detection System
Our computer vision-based anomaly detection system redefines glass manufacturing for NSG, swiftly identifying beading anomalies in real time. Tailored exclusively for NSG, it minimizes material wastage, optimizes operational efficiency, and ensures quality.
The Requirement
NSG Group, a global leader in glass and glazing, faced challenges in detecting anomalies during their glass production processes, particularly beading that led to glass breakage. The company sought a digital solution to swiftly and accurately identify beading occurrences, prevent glass breakage, minimize material wastage, and optimize operational efficiency.
The Solution
LeewayHertz collaborated with NSG Group to develop a robust AI-powered solution tailored for early anomaly detection in glass manufacturing. The solution focused on identifying beading anomalies in real time, enabling timely alerts to operators and proactive intervention. This tailored approach played a crucial role in cutting costs through the reduction of material wastage, streamlining operational efficiency, and elevating the overall product quality for NSG Group.
Features of the Solution
Real-time Anomaly Detection
The solution swiftly identifies beading in live video feeds, allowing for immediate operator intervention during the manufacturing process.
Strategic Adjustment
Enhances functionality by triggering alerts precisely at the fourth second of the beading cycle, ensuring highly accurate detection.
Accurate Anomaly Detection Alert
Provides precise and accurate alerts upon detecting anomalies, minimizing potential false alarms and ensuring timely operator response.
Benefits of the Solution
Responsive Alerts
Accurate anomaly detection alerts provide prompt and precise information, enabling operators to take immediate corrective action.
Operational Optimization
Timely interventions and adjustments improved overall operational efficiency in glass manufacturing.
Quality Enhancement
Ensured a higher standard of product quality, reducing defects and bolstering the company’s reputation.
Proactive Maintenance
Streamlined maintenance and planning processes, reducing unexpected downtime and facilitating improved resource allocation.
Technologies Used to Build the Anomaly Detection System
| IDE: |
 Google Colab VS Code |
| AI Architecture:: |
 Pretrained ResNet18 AlexNet
 Padim
|
| Tools: |
 TensorFlow PyTorch Keras |
