Visual lumber grading is a challenging task for people, requiring extensive knowledge of wood features, grading rules, and ever changing product value requirements that are market driven. At the same time it can be physically demanding and stressful, as grader performance is constantly audited internally and by both regional and national agencies. Consequently, good grader candidates are hard to find, take a long time to train, and are difficult to retain. The last two decades have seen successful application of automated grading machines, driven by advances in industrial vision and machine learning. Local Corvallis company Lucidyne Technologies develops a line of such machines at the forefront of this technology. Deep Learning based AI, sourced by state-of-the-art color, laser, and x-ray sensors, are coupled with advanced optimization algorithms. The results consistently best human performance by a large margin, while extracting the most value from the wood fiber. Rather than being just a lumber grading machine, the scanner is an enterprise system. Continuously collecting data on defect types, reasons for downgrades, and other fine grained details, system reports and other information may be mined to pinpoint improvement areas throughout a mill's various stages.

Patrick Freeman holds a BS in Electrical Engineering from Clemson University and an MS EE from The Ohio State University (Robotics). After working in both semiconductor and PC board design, he joined local Corvallis wood products industry leader Lucidyne Technologies in 1993. Lucidyne's main focus is the design and manufacture of automated lumber grading and optimizing scanners. Patrick has long headed the company's image processing and AI teams, most recently acting as Chief Technology Officer. In his off hours he volunteers as a software mentor for a local high school FRC robotics team, whose side broccoli harvesting robot project is patent pending.