Computer Vision for Asset Inspection (Pipelines, Turbines)

Overview

Learn end-to-end CV for asset integrity: plan safe image capture, build labeled datasets, train detection/segmentation models, and publish explainable inspection results that trigger maintenance actions for pipelines and wind turbines.

Curriculum

• Acquisition plans: drone flight paths, overlap, GSD, lighting; safety & compliance basics.
• Modalities: RGB vs thermal vs multispectral; when each reveals corrosion, leaks, or erosion.
• Labeling workflow: taxonomy, inter-rater agreement, QA, active learning & data versioning.
• Models: object detection (YOLO/Retina), segmentation (UNet/Mask RCNN), transformer-based CV.
• Pipelines: preprocessing, tiling, augmentation, inference-time tricks, post-processing.
• Metrics: precision/recall, IoU, pixel F1, class imbalance & threshold setting.
• Change detection: baselining, defect growth rates, and prioritization scoring.
• Edge & cloud: on-device inference, streaming, bandwidth/latency trade-offs, retries.
• Reporting: heatmaps, defect snapshots, confidence bands, and regulator-ready PDFs.
• Integrations: CMMS/EAM (SAP/Maximo), GIS layers, and work-order auto-creation.

Benefits of Attending

• Faster inspections with consistent, defensible results and fewer site revisits.
• Playbooks for corrosion/crack detection on pipelines and blade damage on turbines.
• Templates for labeling guides, model evaluation, and inspection reports.
• Practical guidance on deploying at the edge and closing the loop with maintenance.

What You Will Learn

By the end of this program, you will be able to:

• Design safe, high-yield image capture plans for pipelines and turbines.
• Build labeled datasets and train CV models for detection & segmentation of defects.
• Run edge inference and triage results with explainability for field teams.
• Publish CMMS-ready reports that convert findings into prioritized work orders.

Practitioner Profile