摘要
竹片缺陷的自动检测成为制约竹制品企业发展的一个瓶颈,为实现竹片缺陷自动检测,本文提出一种柔性神经树的竹片检测算法。该算法首先利用LeNet-5卷积子模型为竹片特征提取器,将新的参数卷积模型提取的特征作为柔性神经树的的输入节点来构建柔性神经树。将该算法应用于竹片缺陷检测能够解决人工神经网络结构的高度依赖性问题,使模型具有较强的鲁棒性。
The automatic defect detection of bamboo chips has become a bottleneck restricting the development of bamboo products enterprises.In order to achieve automatic defect detection of bamboo chips,we propose a flexible neural tree algorithm for bamboo chip detection.Firstly,the LeNet-5 convolution sub model is used as the bamboo feature extractor,and the features extracted by the new parameter convolution model are used as the features of the flexible neural tree for inputting nodes to construct the flexible neural tree.The algorithm of flexible neural tree for bamboo defect detection can solve the problem of high dependence of artificial neural network structure,and make the model have strong robustness.
作者
张雪琼
陈婷婷
蒋丽峰
杨亚蕾
ZHANG Xueqiong;CHEN Tingting;JIANG Lifeng;YANG Yalei(School of Information Science and Engineering,Fujian University of Technology,Fuzhou,China,350118)
出处
《福建电脑》
2020年第8期33-35,共3页
Journal of Fujian Computer
基金
大学生创新创业训练项目(No.S201910388058)资助。
关键词
柔性神经树
竹片缺陷检测
卷积
特征提取
Flexible Neural Tree
Bamboo Defect Detection
Convolution
Feature Extraction