Automated optical inspection solution for manufacturers.
Multi-tenant platform based on deep learning / 2+ year engagement / 6-person team
About the project
Automated optical inspection (AOI) solution enables fast and reliable automated quality inspection of products along the assembly line. Using deep-learning algorithms, the solution can detect even the smallest defects on heterogeneous or reflective surfaces, ensuring high product quality and minimizing the costs of fixing faults.
Product features
Image dataset annotation
Each image can be individually inspected in detail, labeled, and assigned to the specific class. Images can be combined into groups and then used for training and retraining deep learning models.
Deep-learning
​Neural networks are trained in the background and perform object recognition, anomalies and errors detection, annotations and data segmentation.
Role management system
Access to data, projects and various elements within the system is granular and defined for individual users.
Data analysis
Training data and image recognition results are presented in easy-to-read and well-structured dashboards.
Multi-tenancy
Access to the platform can be performed by different tenants. Data is strictly separated to avoid any leaks between clients.
Services
Software Engineering
Technologies
-
Frontend: Angular, Material UI
-
Backend: Nest, TypeORM, MySQL, ZeroMQ
-
Cloud: Kubernetes, Azure, Docker
-
Machine Learning: Python, TensorFlow, Keras, Pytorch
Product Team
6 engineers
How it works
Though the work is still in progress, I’m happy with the results so far. We’ve managed to release a few key features recently and I’m really enjoying the velocity we’re having with Radency. There hasn’t been anything they can’t do.
CTO, Software Development Company
.jpg)
