Knot shape assessment on various species through X-ray CT scanning

International audienceX-ray computer tomography (CT) is the most promising method for non-destructively analysing knottiness in wood logs or beams. Two methods for measuring knot size and location from stacks of CT-images were developed and applied to a wide range of wood samples. The first method aims to provide exhaustive and accurate measurements by manually pointing and tracking the knots within the image stack in view to reconstruct 3D knot shape and arrangement. The method was applied on a set of 125 logs of 17 European species (mainly hardwoods) with the objective to identify species-specific markers based on knot morphology. The second method, which is more oriented toward industrial uses, is based upon an automated algorithm to individualise the knots and output their size and location without any human intervention. Still in development, this method is yet applicable to simple branch architecture patterns, like for softwoods, assuming that moisture content is relatively homogeneous between sapwood and heartwood. Both methods were applied to a set of air-dried beams of Norway spruce and silver fir. The comparison showed that, depending on the beam, 71 to 100 % of the manually measured knots were detected by the automatic method. The R2 values between manual and automated measurements were greater than 0.90 for knot maximal diameter, volume and inclination, and equal to 0.65 for knot length