NativeVision: Machine Vision for Wood Processing
Detect. Act. Improve.
Designed for your figure.
Greater efficiency in your process.
One system, one point of contact.
What is NativeVision?
NativeVision is a machine vision solution from WINTERSTEIGER Solutions for wood quality control. It integrates inspection, decision-making, and follow-up processes into the production line. The system detects and classifies defects such as knots, cracks, resin pockets, or blue stain during ongoing production and transmits the results directly to the machine. The machine then repairs, sorts, or rejects the wood.
With a traditional scanner solution, the scanner takes center stage: The company checks whether it fits into its process and, if necessary, adapts the process accordingly. The inspection results are transmitted via an interface, and the company must handle the downstream processes either on its own or in collaboration with other equipment manufacturers. NativeVision takes the opposite approach: We tailor the scanner to your process, not your process to the scanner. We implement the equipment adaptation and downstream processes ourselves, taking full responsibility for the entire system.
NativeVision at a Glance
Technology
Accuracy
Error Detection
Materials
Use
Interfaces
Detection, decision-making, and action—all in one seamless process.
How NativeVision Works: DETECT. ACT. IMPROVE.
Three steps in a row: Identify problems, take immediate action, and improve yield.
Spot errors like a trained eye
NativeVision captures each workpiece as a 3D image and a color image and evaluates both together. This principle is called Hybrid Vision: Neural networks classify features that are difficult to define in terms of rules, even though wood varies greatly in color and structure. Traditional, rule-based methods measure characteristics that can be unambiguously described geometrically, such as surface area and dimensions. The height profile refines the classification: If the depth indicates it, a branch with a crack becomes a branch with a hole.
- 20+ defect types in softwood, defects starting at 0.25 mm²
- Knots, including black knots, knots with cracks, and knots with holes
- Cracks, resin pockets, blue stain, red streaks, wane, chips, and lamellae that are too thin
- Custom defect catalog, expandable to include dimensional characteristics or other anomalies
- Pre-trained with tens of thousands of images
From Findings to Action in a Single Process
At NativeVision, the follow-up action determines how the inspection is designed. Because WINTERSTEIGER implements the downstream process steps itself, we know what information is needed and what isn’t. For sorting, the quality class is sufficient; for a repair, the system needs the location and extent of each defect. NativeVision provides both, along with the appropriate action—such as sealing a resin pocket with a patch or replacing a knot with a round plug. The WINTERSTEIGER system determines which station performs the repair.
- Results sent directly to the line’s control system (PLC)
- Standard interfaces: OPC UA, PROFINET, EtherNet/IP, and digital I/O
- Repair, sorting, or ejection on the production line
- Repair logic based on product, quality class, and repair technique
- Inspection during ongoing production
Classify Instead of Discard
When sorted by hand, the thresholds are set strictly as a precaution, and flawless material is downgraded or rejected. NativeVision combines your plant’s defined specifications with the experiential knowledge of your employees, which is trained into the models. This ensures that material is evaluated according to consistent standards. After training cycles, the system achieves up to 98% detection and classification accuracy for softwood. Mutually agreed-upon acceptance criteria determine what this means for your project.
- Plant-specific specifications regarding parameters and threshold values, which can be adjusted relatively quickly
- New features via retraining, even when the system is already in operation
- Retraining outside of production; new model deployed only after benchmarking against the current one
- Return to a previous model version is possible
- The images from your production remain at your facility
One system, one point of contact
NativeVision is a joint solution developed by WINTERSTEIGER Solutions and its Danish technology partner, JLI vision. WINTERSTEIGER contributes its expertise in mechanical engineering, process engineering, and woodworking, and integrates the inspection into the production line. In doing so, we build on our own experience with vision systems: The Faulty Spot Eye has been in use in our Timber Repair & Cosmetics systems for over ten years. JLI vision has been active in industrial image processing for more than 40 years, including in critical industries such as the pharmaceutical sector, and provides technology rather than finished products. WINTERSTEIGER Solutions GmbH holds a minority stake in JLI vision.
What this means for you
- A single point of contact for the machine, inspection, and the entire process—from workpiece handling to follow-up actions
- Responsibility for the entire process, not just for a single component
- Cameras, lighting, and sensors that are manufacturer-independent and tailored to your production line
- No off-the-shelf scanners: modular components that are neither undersized nor oversized
Applications of NativeVision in Wood Processing
NativeVision is designed for solid wood processing, for new production lines and as a retrofit solution.
Laminate Testing for Glulam and CLT
Automated Wood Repair
Outlook: New Applications
Machine Vision in WINTERSTEIGER Systems
WINTERSTEIGER has been using machine vision for many years, for example with the Faulty Spot Eye in wood repair and by manufacturers of hardwood flooring, formwork panels, and boards.
New Line or Retrofit: How to Get Your Project Started
Whether it’s a greenfield project or the optimization of an existing line, NativeVision is available for new lines and as a retrofit. Our systems are modular in design, and we can integrate them with systems from other manufacturers via defined interfaces. Whether a retrofit is possible depends primarily on available space: A gap between the conveyors is required for inspecting both the top and bottom surfaces.