Methods for automatic grading of saw logs

The aim of log grading for pricing is to apply an economic value based on volume and log grade. However, unlike many other commodities logs are hard to characterize into different grades. The reasons are: internal properties are hard to decide from the external appearance, the wood properties are often heterogeneous and there are a variety of different grading criterias. In Sweden, as in many other countries, the log grading is based on visual assessment of log characteristics, which can be considered intricate, time consuming and monotonous. Moreover, factors like increasing labour cost, structural changes among sawmills and the introduction of new technology, are all changing the possibilities and demands for the sawmills. New technology is important in order for the sawmill industry to remain competitive and therefore the log grading system for pricing needs to become more flexible and adaptable to the new technology. This thesis investigates two different methods using automatic measuring techniques to grade logs for pricing. In the first method a 3D log scanner is used to estimate log shape and this information is used to grade the log. The accuracy when using 3D log scanning to grade logs, according to the grades in the current Swedish national grading system, was comparable with but somewhat lower than manual grading of logs. Scanner data were used to estimate curvature type and compression wood content. Models based on log scanner data could discriminate between curvature types if the curvature was severe and bow height larger than 0.8 %. The models predicting compression wood were significant even though models using log end information were better. Analysis of whole stems showed that compression wood content was most pronounced near ground level and that stems with basal sweep more often had large amounts of severe compression wood. The second technique used acoustics and the results indicate that the technique has potential for grading logs by the stiffness of the sawn timber. The stiffness criterion was more important for spruce saw logs since spruce lumber is more often used for construction purposes. Using 3D scanning, acoustic methods and other techniques would probably reduce costs and make log grading for pricing more efficient. To enable this, the log grading system needs to be simplified and adapted to the new automatic technologies. Based on the findings in this thesis, a base for a new grading system for pine and spruce is proposed