A machine vision approach for classification and dimensional design of furniture panels using GMM-SVM

作者：Yuan Tian, Zhao Li, Haoxin Li · 发表于：Systems and Soft Computing · 年份：2025 · DOI：10.1016/j.sasc.2025.200358 · 被引用次数：3 · 研究领域：Industrial Vision Systems and Defect Detection、Color perception and design、Color Science and Applications

The categorization and size design of furniture panels have a significant influence on furniture manufacturing and sales. The study presents a classification, identification, detection, and design approach based on machine vision to effectively enhance the quality of furniture manufacture. This method combines Gaussian mixture model (GMM) and support vector machine (Support vector machine, SVM) to effectively identify classification. The findings of this study reveal that the use of various feature parameters and kernel functions affects the support vector machine’s recognition accuracy. Among them, the HSV color space feature parameter set corresponds to the highest recognition accuracy value 0.948. And the average recognition rate associated with the Gaussian kernel function has the greatest value of 0.915. Meanwhile, in the comparative experiments with Bayesian classification algorithm (BC) and artificial neural network algorithm (ANN), the suggested GMM-SVM combination algorithm provides the greatest recognition effect, with the maximum accuracy rate of 0.948. The GMM-SVM method has the shortest recognition time for the same number of samples. In addition, for the plate size design system, the errors obtained in the experiments are small, which can meet the actual needs. The method’s great performance may play a beneficial role in the categorization and design of furniture panels, effectively improving production quality and efficiency.