Finding the Crack Before the Saw Does Rescanning Logs to Protect the Cut Solution

This application covers log and cant optimisation in a sawmill primary breakdown line: scanning a log in three dimensions, detecting splits including their depth, and rescanning the partly cut workpiece to re-optimise the remaining cuts. The applicant is USNR, a major supplier of sawmill machinery and optimisation systems.

The Problem

Modern sawmilling is an optimisation exercise. A log is scanned, software calculates the cut solution that extracts the greatest value from that particular shape, side boards and centre boards are cut accordingly, and the pieces are trimmed to length. It works well and it is the reason yields have improved so much over the past few decades.

The weakness is that the optimisation is calculated once and assumes the world cooperates afterwards. The cut solution presumes the log will be in a particular position when it reaches the saw. If the log shifts after passing the scanner, or is not turned to the correct angle, the solution may not be achievable at all, and boards get cut wrongly. There is no feedback between calculating the plan and executing it.

Splits make this worse. A split running through several cut products can drastically reduce the value recovered from a log. They are also hard to see: a debarked log has a rough outer surface in which a split is difficult to distinguish in an image, and a two dimensional image says nothing about how deep it goes. Mills try to manage this by rotating the log to place the split at a set angle, typically 270 degrees, before cutting. That helps if the rotation is accurate. If it is not, the recovery loss arrives anyway.

What This Invention Does

The disclosure describes a primary breakdown line with two scan zones rather than one, each with its own sensor arrangement and optional sub-zones. The first generates a three dimensional model of the log and detects splits within it, including depth information rather than a surface mark. The second scans the cant), the squared off remainder after the first cuts, generating a rescan 3D cant model.

That second model is what closes the loop. The cant is re-optimised against its actual measured shape and position, so a log that shifted or was turned imprecisely is cut according to what is really there rather than what was planned upstream. The methods claimed cover processing a primary workpiece, generating the 3D log model, detecting splits, generating the rescan cant model and re-optimising the cant.

The disclosed user interfaces are practical rather than incidental. One shows detected splits against an unfurled log model, presenting a cylindrical surface as a flat map so an operator can see where a split runs. Another exposes user selectable optimisation rules, letting a mill weight the solution towards its own product mix and prices rather than accepting a fixed objective.

Key Features

• Two scan zones. Separate sensor arrangements scan the log before breakdown and the cant afterwards.
• 3D log model. A three dimensional model is generated rather than relying on profile measurements alone.
• Split detection with depth. Splits are detected in the model including how far they penetrate, not just where they appear on the surface.
• Rescan cant model. The partly processed workpiece is remodelled after the first cuts.
• Cant re-optimisation. The remaining cut solution is recalculated against the cant’s actual measured shape and position.
• Unfurled split visualisation. An interface presents detected splits on a flattened representation of the log surface.
• User selectable optimisation rules. Operators can adjust the rules the optimiser applies to suit their product mix.

Who Is Behind It

USNR is a long established American manufacturer of sawmill equipment and scanning and optimisation systems, formed through the consolidation of several historic wood processing machinery names and supplying mills across North America and beyond. The named inventors are Douglas G. Strasky, Stephen Doiel, Gerald David Larson, Robert Arnold and Ryan Munion. The application is a divisional of Australian application 2024200556.

Why It Matters

Sawmill economics turn on recovery, the proportion of the log that leaves as saleable lumber rather than chips. Improvements measured in single percentage points are worth pursuing because they apply to every log through the mill, every shift, for the life of the equipment. That is why scanning and optimisation became standard well before comparable automation reached many other processing industries.

The specific gap being closed here is the assumption of perfect execution. Any system that measures once and then commits is exposed to everything that happens between measurement and action, and in a mill that includes a heavy irregular object being turned and conveyed at speed. Rescanning is conceptually simple and mechanically inconvenient, which is presumably why it took until now to appear in this form.

Split detection with depth is the more technically interesting element. Surface features are what cameras see easily; internal structure is what determines value. Inferring depth from what is visible on a rough debarked surface is the kind of problem that has become tractable through better sensing and modelling, and it addresses a defect category that can wipe out the value of an otherwise good log.

Related Concepts

• Sawmill – the facility where this optimisation runs.
• Lumber – the product whose value the cut solution maximises.
• 3D scanning – the measurement technology behind the log and cant models.
• Mathematical optimization – the discipline underlying cut solution calculation.
• Wood drying – the process that produces many of the splits being detected.
• Machine vision – the sensing field applied to defect detection in rough surfaces.

---

AU 2026202075 was published in the Australian Official Journal of Patents on 9 April 2026 and is open for public inspection. Patent applications represent inventions that are sought to be protected and do not necessarily reflect commercially available products.