1. Universal Camera Compatibility
No proprietary lock-in. SightQC runs using pre-existing digital camera systems. Connects to standard industrial setups over GigE Vision or RTSP protocols.
- • Supported: Basler acA series, Hikvision DS-2CD series, Axis M30 series
- • Connectors: PoE GigE, RTSP video streams, USB 3.0 interfaces
Caption: Standard industrial cameras installed on your existing production line.
2. Custom Model Training
Train models customized directly to your unique goods. Supply 50 good and 50 defective images. Our cloud pipeline performs automated augmentation and training.
- • Training engine: Amazon SageMaker custom pipelines
- • Active learning: Factory floor corrections improve models in real-time
Caption: Factory workers monitor quality in real-time.
3. Real-Time Defect Detection
Combines YOLO-based object detection and semantic segmentation to locate defects. An anomaly search engine detects brand new defects that were never trained.
- • Edge Latency: <100ms locally on IoT Greengrass nodes
- • Categories: Scratches, misalignments, color mismatches, missing components, stitching errors
Caption: SightQC catches defects in real-time with 99.2% accuracy.
4. Intelligent Alerting & Rejection
Integrate edge inference outputs directly with rejection relays. Send digital signals directly to pneumatic actuators to discard bad parts.
- • Physical output: GPIO triggers, MQTT messages, and Modbus control packets
- • Mobile alerts: Instant SMS / email notifications on spike triggers
Caption: Real-time alerts on factory floor devices.
5. Defect Analytics Dashboard
Export compliance logs for buyers. Run trend calculations to cross-reference which shifts or supplier batches trigger higher defect flags.
- • Analytics: Trend charts, defect classification pies, and operator shift metrics
- • Exports: Audit-ready PDFs, CSV spreadsheets, and REST JSON data
Caption: Real-time defect analytics help you identify root causes.
6. Multi-Site Management
Manage and track metrics across separate locations. Distribute distinct ML models to separate lines from a unified administrative control panel.
- • Control panel: Secure centralized management dashboard
- • Scaling: Deploy update models remotely across distinct factory locations
Caption: Enterprise-grade architecture built on AWS.