论文部分内容阅读
介绍了目前国内外通过计算机断层扫描技术(CT)结合各种算法对木材中节子部分进行无损检测所取得的研究进展,概述了以灰度阈值法、滤波算法、最大似然法和神经网络法为主的算法识别特点,并对其中应用广泛的神经网络算法进行了对比分析。现有研究表明,运用该技术结合多种算法可实现对原木中节子参数特征的提取与分析,通过算法的不断改进能够提高节子检测的准确率。文中还总结了CT技术在处理节子检测方面存在的主要问题,并对未来趋势进行了展望。
In this paper, the research progress of nondestructive testing of knuckles in wood by computerized tomography (CT) combined with various algorithms at home and abroad is introduced. The methods of gray threshold, filtering, maximum likelihood and neural network France-based algorithm to identify features, and widely used in which a neural network algorithm for comparative analysis. The existing research shows that the technique can be used in combination with a variety of algorithms to extract and analyze the knot parameter features in logs, and the accuracy of knot detection can be improved by the continuous improvement of the algorithm. The article also summarizes the main problems of CT technology in the processing of knot detection, and prospects the future trends.