Artificial intelligence in visual veneer analysis improves on-grade accuracy

A visual veneer analyzer detects sound knots, splits, and other defets in a veneer mat or sheet to determine its grade. The more accurately each defect and their size is recognized, the more accurately the material can be graded for optimal processing and end-use.

Conventional visual defect recognition relies on contrast to analyze defects: the bigger the contrast, the easier the defect is to spot. Some defects are surrounded by wood grain patterns that lower the contrast, making these difficult to spot with conventional cameras. Some defects, such as harvester marks and small sound knots, are so challenging to spot that conventional cameras cannot detect them at all.

To tackle this challenge, Raute has developed visual analyzers that utilize artificial intelligence to spot challenging defects with better accuracy to improve the quality yield.

Modernize your veneer production line with AI powered analyzers  

In principle, any defect that can be detected by humans can also be taught to artificial intelligence. When all defects, including rub marks, harvester marks, or bark and decay, are detected accurately, the veneer sheets can be graded with utmost accuracy, which ultimately results in better end-product quality, raw material recovery, and profit. 

Read more about why and how you should grade veneer sheets in the first place: https://www.raute.com/blog/solutions/why-and-how-should-you-grade-veneer-sheets/ 

For example: typically, conventional systems cannot accurately determine the quantity and size of sound knots. This can result in either: 

• a high number of smaller knots remaining undetected, which means that the veneer sheets will be categorized inaccurately to higher grades. 
• overestimating the size of sound knots, which means that the sheets are inaccurately downgraded. 

“When you modernize your line with an artificial intelligence powered analyzer, you can detect critical defects such as harvester marks as well as sound knots and their sizes with high accuracy. This improves your on-grade accuracy, helps avoid over-processing of veneer sheets, and minimizes financial losses that you have not been able to control before”, explains Tuukka Sorvari, Sales Manager of Analyzers at Raute. 

Accurate grading information allows the existing systems and operators to make better decisions from peeling to panel repairing. 

“AI visual analysis pays back its investment cost through improved end-product quality and better raw material profitability: when you understand the true potential of your raw material at each process phase, you will increase value throughout the mill and boost efficiency in all phases by sending material to the correct destination at the first chance you get”, says Sorvari. 

Example of AI in use

The earlier you use AI in grading veneer sheets, the more your mill profitability improves 

AI visual analyzers can be used in multiple parts of your process to improve profitability - read more about the benefits of AI-powered defect detection in different phases. 

Better grading accuracy in the peeling line improves efficiency in your whole production  

Conventionally, low recovery and bad quality veneer sheets are taken out of the mill’s raw-material flow at the core sheet composer. What if you start picking them out already before the peeling line stacker? 

Decisions made at the peeling line impact the efficiency of all the following process phases. That is why the optimization of quality recovery starts already at the peeling line by understanding the raw material and its features with utmost accuracy.  

Read more about the benefits accurate grading brings at the peeling line: https://www.raute.com/blog/solutions/grading-at-peeling-line-frees-drying-capacity-and-saves-energy/

In veneer peeling, AI helps accurately detect round-up defects such as bark, and defects that can be patched. An AI analyzer brings value through extremely accurate clipping and raw material utilization. When you can optimize clipping and grading already at the peeling line, this improves your quality recovery.  

At the peeling line, you can also utilize built-in patching, composing and drying simulations of Raute Analyzers to simulate raw material usability in the following process phases. These simulations provide you the perfect tools to ensure that the material that is sent from the peeling line onwards is suitable for your developing production needs. 

For example, by using the core composing simulation feature at the peeling line, you can control the veneer recovery level at the core composing process in the future. This allows you to control the veneer inventory volumes to improve your overall process efficiency. 

Example of AI in use 

AI analyzer at the drying line optimizes sheet flow and end-product quality 

If you use veneer sheets as core veneer when they have the potential for face veneer quality, you miss out financially. Poor defect detection can also cause significant losses in profit through unnecessary panel downgrading. 

With conventional machine vision solutions in veneer visual grading, a very common challenge is to separate sound knots, dark knots, loose knots and certain bark defects from each other. If dark heart wood or grain pattern is added into the mixture, the mission is next to impossible.  

With AI, major improvement has been achieved in these previously challenging situations.  
At the drying line, AI can accurately detect defects that need to be patched or composed before further processing, and defects that can be later removed for example at the panel repairing line. Combining AI defect detection with process simulation features, the following processing phases can be accurately selected individually for each veneer sheet. 

For example, you can accurately grade face veneer sheets that don’t need patching already at the dryer, and send them directly to the lay-up line instead of the patching line. This optimizes your sheet flow and improves end-product quality. 

Raute AI analyzers can be retrofitted to any dry grading line. The accurate visual analysis can be combined in the analyzer platform with other features of the veneer - moisture grade, strength properties, waviness and roughness - to make better decisions on how to maximize the potential of your raw material. 

Read more about the benefits of accurate grading in the drying line: https://www.raute.com/blog/solutions/grading-at-the-drying-line-maximizes-profit/ 

Artificial intelligence in composing improves yield 

Composing lines without AI need operators to rip corners from the sheets or mark the defects with fluorescent chalk to make sure the defects are noticed and cut. This manual work is time-consuming and inaccurate, slowing down the process and impacting your output volume.  

Without an accurate analyzer, you end up clipping more defects out than is needed, losing valuable raw material in the process. AI analyzers recognize these low contrast defects without the operator’s manual work, which has a huge positive impact on your yield and capacity in both face and core veneer composing lines. 

Read about our customer Koskisen’s experiences on using AI analyzers on birch veneer at their composing line: https://www.raute.com/references/koskisen-oyj/elevating-plywood-precision-and-productivity-with-ai-driven-veneer-composing/

Patching with AI visual analyzers minimizes veneer downgrading 

Conventional analyzers cannot detect all the defects that should be patched, meaning that the value of the patched veneer decreases. Typically patching line analyzers need to be finetuned to compensate overall inaccuracies in defect detection, causing excessive patching. This in turn causes veneer downgrading, financial losses as well as poor patching efficiency. 

With AI, the level of defect detection is superior, eliminating these challenges. This optimizes line efficiency by reducing unnecessary patching and increases end-product value.  

AI in panel repairing improves panel value 

In panel repairing, even a single undetected defect may cause the whole panel to be downgraded, easily leading to over 50% loss in panel value. 

Improved defect detection accuracy realizes the full potential of the panels: only critical defects are identified and repaired accurately according to their real size. By recognizing the potential panel value already at the repairing line, you can maximize your end-product value. 

AI-enhanced grading brings various benefits for different processes

Modernize your lines with AI-powered Raute analyzers 

Artificial intelligence is part of new Raute analyzers. Each mill is unique, and it might not be feasible to modernize all analyzers at once. Raute experts are available to help you prioritize these opportunities to maximize the benefits for your mill. 

If your current Raute visual analyzers are less than 10 years old, you can send us digital samples. Our AI experts will demonstrate the effect and value of the AI in your production! 

Read more about the mill-level utilization of AI based visual grading to understand how high-end defect detection transforms production and profits: https://www.raute.com/blog/solutions/what-mills-risk-without-high-end-defect-detection/