The Weld Defect Detection sample application demonstrates a powerful fusion of vision and time series analytics to identify welding anomalies. By simultaneously analyzing camera feeds for visual defects and sensor data for abnormal patterns, the application provides a comprehensive assessment of weld quality. Fusion analytics intelligently combines insights from both modalities, applying configurable AND or OR logic to flag anomalous welds. This multi-modal approach enhances detection accuracy, reduces false positives, and enables early identification of potential issues in automated welding processes. The solution is adaptable, supporting various deployment scenarios and integration with existing manufacturing systems.
To see the system requirements and other installation, see the following guides:
- Get Started: Step-by-step guide to getting started with the docker compose deployment of the sample application.
- System Requirements: Hardware and software requirements for running the sample application.
Refer to the How it works section.
- How to build from source and deploy: Guide to build from source and docker compose deployment.
- How to deploy with Helm: Guide for deploying with Helm.
- How to configure MQTT alerts: Guide for configuring the MQTT alerts for the sample app.
- How to update configuration: Guide for updating the configuration.
- Troubleshooting: Troubleshooting information.
- Release Notes: Information on the latest updates, improvements, and bug fixes.