论文部分内容阅读
为改善现有图像位移测量,开发了一种二维位移测量系统。该系统采用3种非线性映射标定方法(BP神经网络、最小二乘(LS)三次曲面拟合、LS与遗传算法(GA)的组合方法(GALS)),基于Microsoft Visual Studio的集成平台。用该系统进行了位移测量实验,并分析了相关影响因素。结果表明:在这3种标定方法中,LS法的速度快、准确性高,综合性能最好;较低像素图像(不小于408×306)和较大的平滑、开运算模板(分别不大于标志点直径的1/2和1/3)有利于获得较好的测量效果。该系统可实现灵活、准确和实时位移测量。
In order to improve the existing image displacement measurement, a two-dimensional displacement measurement system has been developed. The system uses three kinds of nonlinear mapping calibration methods (BP neural network, Least Squares (LS) cubic surface fitting, combination of LS and genetic algorithm (GALS)), Microsoft Visual Studio-based integrated platform. The system was used to measure the displacement and analyze the related factors. The results show that the LS method has the advantages of high speed, high accuracy and best comprehensive performance among the three calibration methods. The lower pixel image (not less than 408 × 306) and the larger smoothing and opening template (no more than 1/2 and 1/3 of the diameter of the marker points are good for obtaining better measurement results. The system enables flexible, accurate and real-time displacement measurements.