Automated visual inspection of wood boards: selection of features for defect classification by a neural network

摘要

Attempts are being made to automate the process of wood sheet grading by using automated visual inspection (AVI). In AVI of wood sheets, much work has been performed on the segmentation and classification stages but relatively little on the extraction of features from segmented images for defect classification purposes. This paper concentrates on feature extraction and presents 32 features potentially suitable for characterizing wood defects. The paper describes a technique for evaluating features and discusses its application to the selection of the features helpful for accurate defect classification by a multi-layer perceptron.