Explosion-Proof Paint Cobot Inspection Pipeline with OpenCV and Fanuc Telemetry
Hazardous paint booth environments require precise cobot path execution combined with real-time visual paint quality inspection. You will create an automated inspection and motion control pipeline using OpenCV and PyTorch to analyze paint thickness uniformity and detect surface drips from Fanuc CRX-10iA/L painting cobots. The vision pipeline must detect paint flaws larger than 0.5mm with 95% precision.
What you are building
Construct an OpenCV and vision-transformer computer vision pipeline that processes paint thickness image frames and correlates defects with Fanuc cobot spatial coordinates.
How work is evaluated
Evaluates paint defect detection precision and cobot spatial coordinate mapping accuracy.
Datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
respray_coordinate_validity
Verifies target respray coordinates remain within reach limits of Fanuc CRX cobot arm.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
defect_detection_precision
Precision of defect classification on test dataset • target: 0.95 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Learning goals
Process high-resolution paint booth optical and thermal camera frames using OpenCV
Train/deploy lightweight defect classification models identifying paint runs, sags, and orange-peel textures
Map dynamic image bounding box coordinates to Fanuc cobot 6-DOF Cartesian workspace points
Measure inspection inference latency to support real-time re-spray trajectory corrections
Resources
Annotated synthetic paint surface optical images featuring sags, drips, and orange-peel flaws paired with Fanuc cobot encoder logs.
Challenge setup
Evaluates paint defect detection precision and cobot spatial coordinate mapping accuracy.
Challenge input
JSON containing image_frame_base64 or pixel array, and cobot_pose_x_y_z
OpenCV
Standard industrial vision processing library.
PyTorch
Standard deep learning framework.
Evaluated output
JSON containing defect_detected boolean, defect_type ("SAG" / "DRIP" / "ORANGE_PEEL"), confidence_score, and respray_target_xyz
- Verifies target respray coordinates remain within reach limits of Fanuc CRX cobot arm.
- Precision of defect classification on test dataset • target: 0.95 • range: 0-1
- Benchmark: Industrial Surface Inspection Benchmark
- Defect Detection Precision target: 0.95
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Shows the first task, up to three tools, and up to three evidence items.
View technical recipe
Configured tools
- PyTorch Lightning · Required
- Captum (PyTorch) · Optional
Evaluation contract
- respray_coordinate_validity · Weight 1
- defect_detection_precision · Weight 1
Metrics
- defect_detection_precision · Target 0.95
Benchmarks
- Industrial Surface Inspection Benchmark
- robotics
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing image_frame_base64 or pixel array, and cobot_pose_x_y_z
Output: JSON containing defect_detected boolean, defect_type ("SAG" / "DRIP" / "ORANGE_PEEL"), confidence_score, and respray_target_xyz
Execution harness: Python
Evaluation tools
- OpenCV · Computer Vision Framework
- PyTorch · Defect Classification Model Engine
Recipe freeze marker
No recipe freeze time is recorded.