Speaker
Description
In the paper, machine learning knowledge is transferred to wood processing industry. А universal Machine Learning system has great potential for the scalability in wood processing industry. In order to increase the yield of high-quality material and reduce the influence of subjective factors in the assessment of the quality of the wood boards, a machine learning model was trained as part of the research. The model was trained on a large set of aspen and black alder board samples, including subjective assessments of the acceptability. This approach ensures that the ML model is able to perceive and standardize different defect compositions that are considered to meet the specified quality requirements. As a result, a more accurate and consistent interpretation of wood defect recognition is achieved, significantly reducing the variation caused by the human factor along with proper feeding mechanism increase the yield of high-quality material. We proved that using ML for wood defect detection allowed to achieve a fully developed and sustainable defect recognition model within 3 months testing period. Identifying wood defects in the production of aspen and black alder boards is both a technologically complex and subjectively influenced process. By combining the assessment of defect intensity, area, structure and location, the model ensures a uniform quality standard throughout production. This reduces unjustified board rejection, improves material utilisation rate and promotes a more stable output of high-quality material