<![CDATA[Blue stain and fungal growth are visual defects on wood surfaces that can reduce aesthetic quality, complicate grading, and in some cases indicate unfavorable storage or moisture conditions. Manual inspection depends on operator experience and is difficult to scale. This article proposes a computer-vision pipeline that combines color-based segmentation with a convolutional neural network (CNN) to detect healthy wood, blue stain, and fungal regions. Color information is transformed from RGB to HSV and YCrCb to exploit hue, saturation, and chrominance differences, followed by thresholding and morphological operations to obtain candidate regions. The candidate regions are cropped into fixed-size patches and classified using a lightweight CNN. The evaluation design uses precision, recall, F1-score, accuracy, Intersection over Union (IoU), and Dice coefficient so that both classification and spatial segmentation can be assessed. Public wood-defect data are used as a reference source for blue-stain samples, while fungal images require project-specific annotation because no experimental fungal dataset is available. This article is a research design; therefore, empirical result values are not yet presented and will be reported after the target dataset is tested.]]>
Copyrights © 2026
