engleski

Features Extraction and Texture Defect Detection of Sawn Wooden Board Images

This paper addresses features extraction from images of sawn wooden board with application in wood surface defect detection. Selection of statistical features is considered. Defects detection in wood images is done by calculating a vector difference between the texture without defects and those with present defects. The method was tested in an ideal case with notable success rate in detecting defects. In addition, experimental analysis is performed on 25 samples with varying number of defects. conclusions on method success is given with discussion on possible improvements.

statistical features ; sawn wooden boards ; image processing ; defect detection ; quality control

nije evidentirano