Predicting Log Type and Knot Size Category Using External Log Shape Data from a 3D Log Scanner

Abstract

The use of new technology such as 3D scanners for measuring log shape at sawmills is increasing. These scanners can provide a detailed model of log shape and evenness of the log mantle area. These data can be used for optimizing yield but also for predicting quality of the sawn goods. This report presents a model for quality - related log features based on data from a 3D log scanner. The model includes routines for development of variables related to four categories of log properties: surface unevenness, log taper, cross - sectional out - of - roundness and straightness. In total, 230 Scots pine (Pinus sylvestris L.) logs are used for developing and validating logistic regression models for sorting logs into classes of log type and knot size. The models are tested on validation data: 97% of the logs were sorted correctly according to log type and 79% were sorted correctly according to knot size.