Handling positioning errors when optimizing sawing of Scots pine and Norway spruce logs using CT scanning

Since computed tomography (CT) X-ray scanning is becoming a reality in sawmills, different studies have been made to establish how well the sawing position of a log can be optimized using CT data. It is also known that positioning errors have an adverse effect on optimization, since the optimization methods used are rather sensitive to positioning errors. To mitigate the effect of positioning errors, a method is proposed in this article that filters results produced by sawing simulation, using a Gaussian filter of a size according to the positioning error. Using these filtered values for optimization, it is possible to retain two percent extra value of the sawn timber, when rotation and offset errors are present, compared to a regular optimization method. A method more robust to positioning errors is more useful in practice, since positioning errors of various magnitudes are always present in sawmills. The main contribution of this paper is, therefore, an optimization method that reduces the effect of positioning errors.