AI Timber Detection, Counting & Dimensioning

We built an AI system that can detect, reconstruct, count, and dimension timber fully automatically. What started as a computer vision challenge turned into a complete end-to-end measurement pipeline.

Timber operations require accurate log counting and dimensioning for inventory management, logistics, and pricing. Manual measurement is labor-intensive, error-prone, and doesn't scale across large stockyards or transport loads.

We developed a multi-stage computer vision pipeline that handles the full workflow from raw image input to structured measurement output.

Instance segmentation to detect and isolate individual logs in stacked configurations

3D reconstruction from multi-view imagery to capture volumetric data

Automated counting with occlusion handling for partially hidden logs

Dimensional extraction (diameter, length) calibrated against known reference points

Custom-trained segmentation models built on PyTorch

3D point cloud processing and reconstruction

Python-based measurement and calibration pipeline

Containerized with Docker for deployment flexibility

The system replaces manual timber measurement with automated, camera-based analysis. It processes images of log stacks and delivers accurate counts and dimensions in seconds, enabling faster inventory turnover and more precise pricing for timber operations.