摘要
以彩色信息图象处理技术为基础 ,围绕光弹性实验数据自动采集这个主题 ,提出用BP神经网络自动确定等色线分数级条纹及其级数的颜色插值法 ,以及全场数字化确定等倾角的五步彩色相移法。典型实验表明 ,上述方法在提高光弹性数据采集自动化程度和减少各种误差因素影响等方面均有所改善 ;进而提高三维光弹性薄切片应力分析结果的精度 ,促进实验应力分析的工程应用 ,扩大光弹性法的应用范围。
On the basis of color imaging system, an automated photoelasticity in white light for determining isochromatics and isoclinics is presented It includes two techniques One is automation of the ordering of color isochromatics based on BP neural network The other is whole field digital determining of isoclinic angles by five step color phase shifting Many limitations and drawbacks, which affect the precision of measurement, were more eliminated and improved The validity of experimental results was checked up with the theoretical exact solutions of a circular disk under diametric compression The results indicate that the techniques can improve the precision of experimental stress analysis in 3 dimentional photoelastic thin slices and extend the application and development of photo mechanics in structural analysis
出处
《机械强度》
CAS
CSCD
北大核心
2002年第2期232-236,共5页
Journal of Mechanical Strength
关键词
彩色图象处理
等色线
等倾线
人工神经网络
相移法
光弹性自动化
Color imaging
Isochromatics
Isoclinics
BP artificial neural network
Phase shifting
Automated photoelasticity
