摘要
金相分析中,粒子识别是重要的一环节,也是具有现实意义的部分,本文结合图像处理技术,将BP神经网络算法融入到金相粒子识别当中,通过图像的预处理、图像分割、特征提取和BP训练来完成这一目标,通过实验和算法改进,最终识别验证有效,总体识别准确率大于90%。
particle identification is an important portion in Metallographic analysis,and it is very common useful.This article take the BP neural network algorithm practice into the analysis of metallographic with image process technology to get the eventually identification result with an effective accuracy than ninety percent,which through image recognition pretreatment,image segmentation,feature extraction and BP training to reach.
出处
《电子世界》
2018年第8期16-17,20,共3页
Electronics World
关键词
金相分析
图像处理
BP神经网络
Metallographic analysis
hnage processing
BP neural network
