机器视觉在木材选色中的应用

Application of machine vision in wood color selection

在木材加工领域,木材表面颜色是否协调对最终产品的经济效益有重要的影响。木材由于天然生长的缘故,表面颜色丰富,人工分拣困难。为对木材颜色进行分类,针对该场景提出使用无监督学习算法,即聚类算法。首先对木材图片进行轮廓检测,数据清洗排除异常样本,然后在RGB色彩空间上进行特征提取,最终经过K-means算法对样本的特征聚类。实验结果表明:在木材颜色分类场景下使用聚类算法可有效减少人工标注的成本,算法运行效率高,具有重要的应用价值。

In the field of wood processing, the color coordination of wood surface has an important influence on the economic benefit of products.Due to the natural growth of wood, the surface is rich in color and difficult to sort manually.In order to classify wood colors, an unsupervised learning algorithm called clustering algorithm was proposed for this scene.Firstly, contour detection is carried out for wood pictures, data cleaning is used to eliminate abnormal samples, and then feature extraction is performed on in RGB color space.Finally, K-means algorithm is used to cluster the features of samples.The experimental results show that the clustering algorithm can effectively reduce the cost of manual labeling in the wood color classification scene, and the algorithm has high efficiency and practical significance in engineering.