Three-dimensional inversion of knot defects recognition in timber cutting

Abstract The comprehensive utilization of wood is the main goal log cutting, but knot defects increase difficulty rationally optimizing cutting. Due to lack real shape data in logs, it difficult for detection methods establish a correlation between signal and defect morphology. An image-processing method proposed inversion based on distance regularized level set segmentation (DRLSE) spatial vertex clustering, with existing relative board position log, an model established. First, edges top bottom images boards are extracted by DRLSE ellipse fitting, major axes ellipses made coplanar angle correction; second, coordinate points form straight line; third, solve intersection dispersion lines axis plane, K-medoids clustering used locate vertex. Finally, large ellipse, 3D cone constructed which can be invert knots board. experiment was conducted ten defective larch boards, experimental results showed that this accurately shapes solid advantages low cost easy operation.