A New Method for Analyzing Defects in Veneer Images: Hypothesis Testing Based on Gaussian Mixture Decision Function

Abstract

Accurate detection of defects plays a vital role in the wood industry due to the direct relationship between the quality and price of wood products. To this aim, in this paper, we introduce a new method in which we first use hypothesis testing to distinguish between wood defects and clear wood. In the proposed scheme, firstly the natural pattern of veneer is removed by applying morphological enhancement, and in the second step, the probable defects are estimated by a decision function based on the Gaussian mixture concept. The performance of the proposed algorithm is evaluated on a data set of veneer images containing several types of surface defects. The results demonstrate that the proposed method extracts the defects approximately 8.2% better than its alternatives, in parallel with decreasing false detections by approximately 7.3%. The results obtained also show the considerable improvements in Accuracy and Precision of the proposed method compared to other examined methods, especially when a high detection rate (i.e. at least 90%) is desired.