Human lumber graders are only 48-75% accurate but grade 90%+ of US hardwood

Hardwood lumber in the United States is overwhelmingly graded by human inspectors who visually assess each board on a moving production line, identifying defects and calculating clear-cutting areas according to National Hardwood Lumber Association (NHLA) rules. Studies show these inspectors achieve only 48% to 75% accuracy. This matters enormously because grade determines price — the difference between FAS (First and Seconds) and #1 Common can be $500+/MBF (thousand board feet). A mill processing 50,000 board feet per day with a 60% grading accuracy is misgrading 20,000 board feet daily, creating disputes with buyers, returns, and either lost revenue from undergrading or customer churn from overgrading. The causes are well-understood: production line speed forces split-second decisions, the mental math of calculating cutting units is error-prone under time pressure, and inspector fatigue from monotonous repetitive work degrades accuracy through each shift. Automated grading systems using multi-sensor scanning (color, X-ray, 3D laser, grain deviation) achieve 92%+ accuracy, but adoption remains low because a single automated grading line costs $1-3M, which only the largest mills can justify when the median US hardwood sawmill processes under 5 MMBF/year.