一种基于先验模板的图像面积特征测量方法

A Novel Method of Measuring Area of Image Feature under Automatic Microscopic Manipulation Based on NCCO

针对显微视觉中需要反复自动聚焦与自动照明调节以获取高质量图像,降低了微操作效率,提出一种图像特征测量方法.该方法建立在二值图像标准互相关计算模型的基础上.首先,通过设计特殊模板图像,简化该计算模型;通过反推该计算模型,可以在不作预处理基础上,得到目标图像的面积特征参量.利用误差补偿原理,设计两种模板求解两次面积参量,而后求平均值,提高了算法对比度及鲁棒性.强噪音仿真图像实验结果表明:改进方法的测量正确度比常规方法高约20倍;强离焦退化仿真实验表明:改进方法的测量正确度可提高一个数量级.实际图像实验中,改进方法可以在严重离焦以及弱照明条件下有效对图像面积特征进行测量,测量精密度比常规方法高3～5倍.结果表明,该方法在低质量图像面积特征测量中具有实用价值.

A method to measure area of image feature under tough noise and badly defoucsing was introduced. Firstly, a probability distribution based formula computing NCCO （Normalized Cross Correlation Operator） with two binary images was proposed. Some special template imags were disigned. Accordingly,mathematic models of these special geometry image features＇ similar functions were derived from and proved. Thus the area of target images can be derived from the model. An improved method of computing average under two different templates, was presented to achieve higher precision and robustness. The results of simulating image experiments show that the improved method of measurement correctness than conventional methods to raise about 10 times. Finally, the practical experiments were conducted on a hybrid microassembly workcell. Experimental results show that the improved method of measurement precision than conventional methods to raise about 3-5 times. These results show that the NCCO based measuring algorithm is a method with practical value.