【作者】 汪宏昇； 夏奇； 周明才； 史铁林；

【Author】 WANG Hong-sheng, XIA Qi, ZHOU Ming-cai, SHI Tie-lin (Mechanical Electronic and Information Engineering Department, Huazhong University of Science and Technology, Wuhan 430074, China)

【机构】 华中科技大学机械学院机电信息工程系；

【摘要】 研究了机器视觉中图像定位技术的原理,分析了相关法、重采样和曲面拟合的方法,对重采样进行如下改进: 对模板进行基于立方Hermite的智能图像重采样算法,锐化图像高频信息,平滑低频信息．有效避免了图像模糊．增强了目标图像的对比度和清晰度,从而提高了重采样的亚象素的定位精度;结合相关法、重采样和曲面拟合法提出一种亚象素级的快速匹配算法,提高算法计算速度,并实现0．2个亚象素的匹配精度,满足实际定位要求。试验结果表明了该算法在图像模糊或图像有一定角度的畸变条件下的有效性、精确性和鲁棒性。更多还原

【Abstract】 The principle of the image localization problem in machine vision system was discussed. The resampling, fitting and matching by correlation methods’ characteristics were analyzed. The improvement was made by using the intelligent interpolation method based on cubic hermite to obtain high accuracy. A new algorithm based on fast image correlation matching was proposed by combining the resampling and fitting method and the accuracy of localization was redized within 0. 2 pixel. Experiment results on some real images and the corresponding simulated ones were given to demonstrate the algorithm’ s efficiency, accuracy and robustness under the conditions of blur and aberration.更多还原