How Hindustan Pencils Achieved 99% Accurate Wooden Slat Inspection with Jidoka

Jidoka's Approach

‍

Multi-Camera Vision Architecture

Jidoka deployed a four-camera inspection system to ensure complete surface coverage:

• Top Camera: Captures high-resolution images of the top surface to detect cracks, knots, chips and pin holes.
• Bottom Camera: Captures multiple images (15 mm field of view per frame) as the slat passes through the inspection gap. The system stitches images together to create a full bottom-face view.
• Side Cameras (2): Inspect both edges to detect bends and structural irregularities.

The wooden slat field of view spans up to 200 mm × 100 mm, ensuring comprehensive inspection of critical surfaces.

AI-Powered Defect Detection

• All captured images and sensor inputs are processed through Jidoka’s deep-learning Kompass engine.
• The system automatically detects and classifies: cracks, knots, chips,pin holes, bends, width & length deviations, thickness variation
• Defects are categorized into critical, major and minor with predefined thresholds, enabling standardized decision-making without human subjectivity.

‍

PLC Integration & Automated Rejection

• Image capture is triggered by Jidoka’s PLC in sync with inspection cycles.
• Inspection results (OK/NOK) are communicated in real-time to the machine.
• Eight reject gates ensure accurate segregation of defective slats.
• Electrical interlocks integrate seamlessly with the conveyor and station controls.
• OK parts continue downstream without interruption.

This ensures zero inspection downtime while maintaining continuous production flow.