Optimizing Lumber Yield with 3D CT Scanners and AI

Modern sawmills maximize the value of every harvested tree by combining 3D X-ray computed tomography (CT) log scanners with artificial intelligence (AI). This article provides an overview of how mills scan the inside of raw logs to map internal defects, use AI algorithms to calculate the ideal cutting pattern around knot geometries, and automatically rotate the tree trunk on the sawline to extract the highest possible yield of structural-grade lumber.

3D X-Ray Computed Tomography Scanning

Before a log touches a blade, it passes through a ring-shaped 3D X-ray CT scanner at high conveyor speeds. Just like a medical CT scanner, an X-ray source revolves around the tree trunk, taking hundreds of cross-sectional density measurements per second.

Because knots, rot, pitch pockets, splits, and normal clear wood all have different densities, they absorb X-rays at different rates. The scanner compiles these cross-sectional slices into a continuous, millimeter-accurate 3D digital twin of the log, revealing both its external shape and internal structure.

Mapping Internal Knot Geometry

Knots are the primary factor that reduces the structural strength and commercial grade of dimensional lumber. The CT scanner's software isolates the dense wood of the knots from the clear heartwood and sapwood, mapping the exact 3D geometry of every branch stub buried inside the trunk. The system identifies:

  • Knot location and angle: Where the knot begins at the pith (center) and how it angles outward toward the bark.
  • Knot size: The exact diameter and volume of each knot along its entire length.
  • Knot condition: Whether the knot is intergrown (tight/sound) or encased (dead/loose), as loose knots severely downgrade structural lumber.

AI-Driven Cutting and Rotation Optimization

Once the internal knot map is generated, knot-geometry optimization AI takes over. In a fraction of a second, the AI runs thousands of virtual sawing simulations on the 3D digital log.

The AI is connected to the sawmill’s current pricing tables and structural grading rules. It tests how the log would break down into boards at every possible degree of rotation (from 0 to 360 degrees) and varied horizontal offsets. The algorithm aims to place the saw kerfs (the path of the blades) directly through the worst defects or position the cuts so that large knots are relegated to the edges of boards, where they can be trimmed off. This preserves clear, knot-free wood in the center of the main structural boards, upgrading them to premium structural grades.

Automated Log Rotation and Breakdown

After the AI selects the single simulation that yields the highest total financial value and structural grade, it sends mechanical instructions to the primary breakdown machine (the headrig or quad-band saw).

As the physical log approaches the saws, automated turners grip the trunk and rotate it to the exact degree dictated by the AI. Simultaneously, servo-hydraulic positioners shift the log horizontally to align with the optimized cutting axis, accounting for any natural curve (sweep) in the trunk. The log is then driven through the saws, turning the virtual optimization into physical, high-grade structural lumber.