Determination of species of some wood veneers using machine vision

The veneer industry is widely used in many countries of the world. Properties such as obtaining curved surfaces, concealing defects, usability in different designs, and patterns can be counted as the main advantages of veneers compared to solid wood. Although advanced production and quality control systems have developed in coating production, the number of countries where traditional production methods are used is quite high. The biggest problem in transition to advanced systems is investment costs. Therefore, companies allocate low budgets for growth. This study carried out machine vision estimation of different types of veneers obtained by the quarter‐cutting and rotary methods. Fifteen different veneers samples were used in the study. Over 923 features were extracted from 75 veneer images to determine the selected features. Artificial neural network and decision tree techniques were used as decision‐making algorithms. It was determined that the artificial neural network made predictions with higher accuracy in estimating the veneer type. Rays, annual rings, and light‐dark color contrast were among the parameters effective in machine vision prediction. The importance of data is increasing in businesses that are undergoing digital transformation and automation. Today, data are expressed not only in numbers or texts, but also in features extracted from images. With this study, numerical features were extracted from the veneer images and species predictions were made. It was determined that among the algorithms used, the artificial neural network yielded more accurate results than the decision tree technique.