Automation of lathe checks detection using deep learning

Abstract Automated detection of lathe checks in wood veneers presents significant challenges due to the variability of lathe checks characteristics and the natural properties of wood. This study explores the use of two convolutional neural networks (U-Net architecture) to enhance the precision and efficiency of lathe checks detection in poplar veneers. Two U-Nets are employed sequentially to map lathe checks using semantic segmentation, followed by post-processing to denoise these mappings and extract lathe checks characteristics. The first U-Net, used for lathe checks detection, demonstrated strong performance in predicting crack presence, with precision and recall scores of 0.822 and 0.835, respectively. The second U-Net, used for lathe checks connection, further refined these predictions by linking disjointed lathe checks segments, thereby improving the overall lathe checks mapping process. Comparative analysis with manual methods revealed comparable or superior performance of the automated approach, especially for shallow lathe checks.