收录:
摘要:
A 2-D image-processing displacement-measuring system was developed to provide improved measurements. The system uses nonlinear mapping calibration algorithms, such as a back propagation (BP) artificial neural network(ANN) method, least square (LS) method, and an combined algorithm (GA-LS) from the genetic algorithm (GA) and LS, in a Microsoft Visual Studio integrated development environment. The system was used to analyze the factors influencing displacement measurements. The results show that the cubic surface fitting based on the LS method is the best among the 3 calibration algorithms. Better measurements can be obtained with a smaller image (but larger than 408 × 306 pixels), larger templates for smoothing (but less than 1/2 of a marking point diameter) and for opening operations (but less than 2/3 of the marking point diameter). The measuring system is flexible and accurate and can be used in real-time.
关键词:
通讯作者信息:
电子邮件地址: