基于Halcon多层感知机的木材纹理识别  

Wood Texture Recognition Based on Halcon Multi-layer Perceptron

作　　者：孙静[1] SUN Jing(College of intelligent manufacturing engineering,Liming Vocational University,Quanzhou 362000,China)

机构地区：[1]黎明职业大学智能制造工程学院,福建泉州362000

出　　处：《唐山师范学院学报》2023年第6期74-77,133,共5页Journal of Tangshan Normal University

基　　金：2020年福建省中青年教师教育科研项目(JAT201308);黎明职业大学科研团队(LMTD202001)。

摘　　要：为解决当前木材分类受人为因素影响大、识别精度低等问题,依据木材显著的纹理特性提出了一种通过灰度共生矩阵和灰度直方图法提取木板纹理图像特征,采用多层感知(MLP)神经网络算法进行训练学习、识别分类的方法。算法在Halcon视觉软件环境下实现。实验结果显示,该方法对不同木板纹理图像识别效率达到86.7%,具有良好的应用价值。In order to solve the problem of manual visual selection in wood classification at present,which are greatly affected by human factors and low recognition accuracy,a method of extracting wood texture image features through gray level co-occurrence matrix and gray level histogram is proposed according to the remarkable texture characteristics of wood,and multi-layer perception(MLP)neural network algorithm is used for training,learning,recognition and classification.The algorithm is implemented in Halcon vision software environment.The experimental result shows that the recognition efficiency of this method for different wood texture images is 88.33%.So this method has good application value.

关 键 词：纹理识别 灰度共生矩阵 多层感知 HALCON 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]