Machine Vision Technology the Forest for Products Industry

W hen a tree is felled, a process begins to maximize the value of that wood. The log is taken to a sawmill, where someone decides whether the wood is more valuable as lumber, veneer, or chips. If the log is made into lumber, boards cut from it must be edged and trimmed, a process that requires someone to decide how to trim off defective parts and make the board as valuable as possible. Then someone must examine the board and give it a grade—based on the quality of the wood and presence of defects—and this determines the board’s selling price. Finally, at the rough mill, someone cuts the lumber again to produce defectfree dimension parts. Much of this process is based on subjective decisions and incomplete information. When a sawyer trims a board, he knows only what he sees on the board’s surface. As the board passes through the mill, decisions must be made quickly, and sometimes a board is trimmed or cut in a less than desirable way. Finally, when the board is graded, the grader makes decisions on the value of the board based only on what he can infer from an inspection. Since this is a subjective process, two people may assign different grades to the same board. In the forest products industry, workers’ decisions directly affect the quality and yield of the wood products created. Unfortunately, the process is slow and subject to error. Lumber is often improperly cut and trimmed, and boards are frequently incorrectly graded. 1 Furthermore, this manual process doesn’t lend itself to the development of automated sawmill operations. Clearly, machine vision technology, with its ability to provide more complete information, could find a place in the forest products industry.