Ultrasonic defect detection in wooden pallet parts for quality sorting

Millions of wooden pallets are discarded annually due to damage or because their low cost makes them readily disposable. Higher quality wooden pallets, however, can be built from high quality deckboards and stringers, and have a much longer life cycle and a lower cost per trip. The long- term goal of this project is to develop an automated pallet part inspection system to sort pallet parts according to grade. Ultrasonic time of flight (TOF) measurements in a pitch-catch arrangement are being used to distinguish types of defects, including knots, decay, cross grain, and voids, from clear wood. Rolling transducers of 3 different frequencies have been used to collect measurements on four oak deckboards of 1/2 inch thickness. Ultrasonic C-scans taken on a 1/2 inch by 1/2 inch grid indicate that TOF with 84 KHz transducers can be used to partially distinguish between several deckboard features and clear wood. Nevertheless, future application of these results to defect detection must not be limited to single, pixel value classification, but must include pixel neighborhoods with textural information.