top of page

AI Needs to Be on Grade

Writer: NHLA
NHLA
Nov 9
3 min read

The NHLA is getting involved with AI graders as part of the $500,000 Wood Innovations Grant it received from the U.S. Forest Service (USFS). The six participating AI grader equipment manufacturers need help ensuring the accuracy of their defect annotations for each new species and its regional variations when they set up a line. The NHLA National Inspectors, under the direction of the Chief Inspector, are the most qualified candidates to annotate images of hardwood lumber for defects that AI can then use to grade lumber. NHLA will provide annotation services, training, and quality control to help ensure that AI graders accurately identify defects and apply the grading rules. This will bring many opportunities for new scrutiny of the rules and how they are keeping pace with technological advancements and a changing resource landscape.


image of nhla grading rules books in english, french and spanish

Simply put, the NHLA Rules for the Measurement & Inspection of Hardwood & Cypress must adapt to technology and resources, or they become obsolete. Or, as I have heard some rules called, “A fallback, in case of a claim.” I will say unequivocally that the FAS wane rule was not created as a guideline. Paragraph 57 states, “Wane shall not exceed on either edge of the piece over one-half the length in the aggregate.” If it has become more of a guideline today, it is because manufacturing technology has changed, and the rule has not kept pace. (Wane is not a resource change; trees still grow round, and lumber is still rectangular.) As one member rhetorically asked and then told me, “Where can you put wane in any finished product? With the small exception of a live edge, it is not wanted.”


I am not here to tell you what the rules should be. I am only here to coordinate and ensure publication of the NHLA Rules. The Chief Inspector’s role is to guide lumber inspectors in measuring and inspecting hardwood and cypress.


The NHLA is the keeper of the hardwood grading rules. However, NHLA National Inspectors have less work as the number of hardwood producers and manufacturers in North America declines. There are fewer claims because it is often cheaper to settle than to have a National Inspector travel to the site and grade the hardwood lumber. Proprietary grades and in-house training also reduce the grading and instruction performed by the National Inspectors. The declines in the number of students taking the full eight-week course and the number of inspectors working at mills are not going to reverse on their own, and the NHLA must adapt.


AI grading may change the number of graders mills need and the training those employees require. but may increase the number of disputes and inspections. That increase in disputes is what we are trying to prevent. Many mills in the USA are looking at or have already installed an AI grader. They need help ensuring the accuracy of their grading, which depends on accurate defect annotations and correct application of the rules. Calculating surface measure and cutting units is just math that a computer can do quickly. The computer can run through many possible cutting combinations to determine the grade, best yield, and optimal cuts. The reintroduction of hardwood structural grades will also require additional training and programming of the rules, like softwood grading. The ability of AI graders to evaluate lumber for yield or structural grading will become critical to optimizing and improving the profitability of the hardwood industry.


The NHLA currently has five National Inspectors who work large territories and travel extensively. The National Inspectors are the most qualified candidates to annotate images of hardwood lumber captured by AI grading systems. This grant funding will allow the NHLA to hire additional National Inspectors to perform other inspection duties, reducing their coverage areas and bringing them closer to mills in the Midwest, Northeast, South, and West. It is anticipated that the National Inspectors will annotate images, on average, one day per week. Each will work with a specific AI equipment company and help the mills using that company’s AI grader.


This is an exciting step for the NHLA and is exactly what we need to do to move forward. It will bring changes to both the inspection work and the rules. But this is for the best. We may approach AI from different sides of the industry, but here is our one common ground: whether lumber is graded by a person or a machine, we can all agree it needs to be on grade.


Dallin Brooks

NHLA Executive Director

FedEx
FedEx
bottom of page