摘要
提出一种利用BP神经网络获取平面光学元件面形偏差的方法。该方法先确定一幅干涉条纹图像测试区域中条纹弯曲量和条纹间距,然后对BP神经网络进行训练,最后获取被测平面光学元件的光圈数。通过与Zygo激光干涉仪的测量结果进行比较,发现两者的测量结果相吻合。实验结果表明,该方法不仅可以增强处理干涉条纹图像的适应性,而且可以提高测量精度。
A method of obtaining the surface form deviation of the plane optical clement by using BP neural network is presented. The curvatures of the interference fringes and the values of the fringe spacing of tested zones in the interferogram are determined, and the BP neural network is trained to obtain the value of light ring of the plane optical element. It is found that the test results obtained by Zygo interferometer are consistent with the test results obtained by the proposed method. Experimental results show that the method can enhance the adaptation capability of processing the interferograms and improve the measuring precision of the plane optical element.
出处
《电子科技大学学报》
EI
CAS
CSCD
北大核心
2010年第4期581-584,622,共5页
Journal of University of Electronic Science and Technology of China
基金
上海市教委重点学科建设项目(J51801)
上海第二工业大学校基金(QD209010)的资助
关键词
BP神经网络
数字图像处理
条纹分析
测量
BP neural network
digital image processing
fringe analysis
measurement
