Optical inspection, grading and jam detection for sawmills and wood processing
- Detects logs that roll over, jam, or get stuck on log turners and production lines, in real time.
- Automatically stops or signals the line before a jam causes equipment damage.
- Works with your existing CCTV cameras and VMS system.
- Automates wood grading and quality assurance on external defects and characteristics, with internal defect analysis (rot, cavities) available where CT or X-ray imaging is in use.
- Line speeds of 200+ metres per minute. Scalable with processing hardware.
- Turnkey delivery includes cameras, sensors, cabling, and rigging based on your needs.
Minimize downtime and reduce incidents
Accurately automate wood grading
Cost effective even for small sawmills
Solutions
Log Jam & Rollover Detection
Detects a log that rolls over, jams, or gets stuck on a log turner or production line, and triggers an automatic line stop or operator alert before it causes damage. Deployable on your existing CCTV and camera infrastructure.
Wood Grading & Quality Assurance
Real-time assessment, grading and sorting based on external defects and characteristics: knots, cracks, wane, warp, and discoloration.
CT or X-ray imaging feeds can be used to detect internal defects such as rot or cavities and sort accordingly.
Jam Detection
On most log turners, log decks and conveyors, there’s no automation watching for a rollover or a stuck log. An operator has to notice and react, and by the time they do, the log is often already jammed against equipment or backing up the line behind it.
MarshallAI watches the line continuously and flags a rollover or stall the instant it happens, stopping the line automatically or alerting the operator before logs cause equipment damage or reach a point where someone could get hurt.
Accurate Grading
Grading raw, rough-sawn lumber is a structural and dimensional check: catching wane, knots, splits and checks, and warp (bow, crook, cup, and twist) to sort each board into the correct structural or dimensional grade straight off the saw.
Grading planed, surfaced (S4S) lumber is a different job. Once the face is smooth, appearance-grade characteristics like mineral streak, pitch pockets, planer snipe, and color and grain matching become what determines value. MarshallAI applies the right detection criteria for each stage, at line speed, on every board.
Fast Deployment
MarshallAI deploys on the CCTV and camera infrastructure you already have, so there’s no new hardware to wait on to get started. Every model is trained specifically for your mill’s layout, lighting, and species mix, not a generic pre-set.
The vision pipeline scales with GPU and parallel processing, keeping pace with line speeds of 200+ metres per minute, well above what most industrial resaws and edgers run today.
Capabilities used
-
Color Detection & Verification
Identifies and verifies color against a defined reference, catching mismatched product, print, packaging or paint color.
-
Anomaly & Defect Detection
Flags trained deviations including damage, contamination, or irregularities that were not explicitly programmed in advance.
-
OCR & Code Reading
Reads and verifies printed or embossed text, including best before dates, batch codes, and serial numbers, even on curved, metal or low contrast surfaces. 1D barcodes, and 2D codes such as DataMatrix and QR.
-
Object Counting & Enumeration
Counts units, parts, or occurrences within a frame or across a production run.
-
Real-Time Actuation & Control
Triggers a physical or digital response, such as a reject arm, gate, or PLC signal, automatically from a detection event. Human-in-the-loop control also possible.
