Automated grading, upgrading, and cuttings prediction of surfaced dry hardwood lumber
This paper concerns the scanning, sawing, and grading of kiln-dried hardwood lumber. A prototype system is described that uses laser sources and a video camera to scan boards. The system automatically detects defects and wane, searches for optimal sawing solutions, and then estimates the grades of the boards that would result. The goal is to derive maximum commercial value based on current market prices. This paper presents the results of a recent empirical test in which the system 's grading decisions are compared with those assigned by a human expert. We also assess the potential of cuttings from the lumber by board grade. The test involved 86 yellow poplar boards and 90 red oak boards. The automated system assigned higher grades for 17% of the boards, and it assigned lower grades for 43% of the boards. The main cause of disagreement was the presence of stains on the board, both natural and mechanical, which were occasionally classified by the scanning system as defects. The system also recommended additional edging or trimming on 42% of the boards to increase the grade and value of the boards. Overall, the automated system performed well on typical cases of planed and dried boards.