Wood veneer surface manufacturing defects —prevalence in Malaysian industry and human baseline defect detection performance

Abstract

Wood products are perceived as premium products. Therefore, visible surface defects are undesirable. The current defect detection in wood products is by manual visual inspection. There is scant research data available on the defects plaguing the downstream wood industry. This paper determined the extent of such defects in assembled wooden veneered interior doors produced in a Malaysian manufacturing plant, focusing on American red oak (Quercus spp.), yellow poplar (Liriodendron tulipifera) and maple (Acer spp.) species. Industrial random sampling defect data was classified into seven defect categories. Pareto analysis showed that handling defects--particularly scratches/dents and knife marks--were the most prevalent, constituting 30% of all defects. The relationship between human ocular physiology and defect detection ability was tested using SPARCS (Spaeth/Richman Contrast Sensitivity) methodology, which was found to be a good lowcontrast ability predictor. Several common errors causing false positives were also identified. Comparisons using statistical t-tests between industry personnel and non-experts, and between genders showed that there was no difference in detection performance. In conclusion, human fallibility was the main cause of failure in detecting defects, particularly those with low contrast. Human behavioural results gathered in this study can be utilised as benchmarks for future automation studies.

Improved Faster R-CNN Based Surface Defect Detection Algorithm for Plates

“…There are also artificial board defect recognition methods based on region screening segmentation and random forest, artificial board surface defect detection image adaptive fast threshold segmentation algorithm, and artificial board surface image defect extraction method based on grayscale cogeneration matrix and hierarchical clustering. All of these methods require manual selection of certain key parameters or production features of the artificial board surface and have the disadvantages of low generalization ability and shallow feature level [ 14 ].…”

Colour sorting of red oak, yellow poplar and maple veneers using self-organizing map: comparisons between different camera genres

“…In colour studies, oak (Quercus spp.) appeared to be the most popular specie, which correlates with its popularity as the most popular hardwood specie in the Western hemisphere, particularly Europe (Krackler et al, 2011) and the United States (Barbu & Tudor, 2021;Tan & Ng, 2019). Several others focus on high value species such as teak (Tectona grandis) and cherry (Prunus spp.…”