LabWindows/CVI环境下散焦测距神经网络补偿器  

Neural network compensator for depth from defocus under LabWindows/CVI environment

作　　者：栾晓明[1] 王鑫 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001 [2]上海铭源数康生物科技有限公司,上海201000

出　　处：《应用科技》2007年第8期1-4,共4页Applied Science and Technology

摘　　要：应用矩不变原理的散焦测距方法,针对由于光学畸变等原因而存在的非线性误差,采用BP神经网络补偿器对检测结果进行了一定补偿.提出了一种在LabWindows/CVI环境下应用ActiveX技术与MATLAB混合编程的方法来实现BP神经网络补偿,并通过训练与学习证明了补偿器的有效性.Since the depth formula for the depth from defocus using moment-preserving principle may yield nonlinear errors owing to optical distortion, BP algorithm of artificial neural network is employed to compensate for errors resulting from the depth formula. The compensation is achieved by joint use of ActiveX technology and MATLAB programming running on environment of LabWindows/CVI . The results from the training and learning show that this compensator is effective.

关 键 词：神经网络 散焦测距 LABWINDOWS/CVI 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]