Knot detection in computed tomography images of partially dried Jack pine (Pinus banksiana Lamb.) and white spruce (Picea glauca (Moench) Voss) logs

X-ray computed tomography (CT) of logs means possibilities for optimizing breakdown in sawmills. This depends on accurate detection of knots to assess internal quality. However, as logs are stored in the log yard they dry to a certain extent, and this drying affects the density variation in the log, and therefore the X-ray images. For this reason, it is hypothetically difficult to detect log features in partially dried logs using X-ray CT. The objective of this research was to investigate the effect of drying on knot detection in Jack pine (Pinus banksiana Lamb.) and white spruce (Picea glauca (Moench) Voss) logs from New Brunswick, Canada. An automatic knot detection algorithm was compared to manual measurements for this purpose, and the results show that knot detection was clearly affected by partial drying. Because dried heartwood and sapwood have similar densities, the algorithm had difficulties detecting the heartwood-sapwood border. Based on how well the heartwood-sapwood border was detected, it was statistically possible to sort logs into two groups: 1) Low knot detection rate, and 2) High knot detection rate. In that way, a decision can be made whether or not to trust the knot models obtained from CT scanning. Therefore, logs that are partially dried out and fall in the low knot detection rate should be handled cautiously because the optimization results based on CT knot detection cannot be fully trusted. Sawing of these logs could be optimized using only their outer shape, ignoring internal quality. Similarly, only logs having a regular heartwood shape should be used when scanning logs for research purposes or in databases of CT scanned logs. Finally, a larger knot detection rate was obtained for Jack pine. This could have been facilitated by the fact that pine trees usually have larger but less numerous knots than spruce trees