摘要
传统的归一化互相关(NCC)匹配算法基于灰度实现,存在运算量大,忽略图像特征信息、无法匹配旋转图像等问题,因此研究了一种基于OpenCV实现的结合图像特征和灰度信息的Edge-NCC模板匹配算法。该算法在图像金字塔顶层利用边缘点进行相似度计算,计算量可减少(1-z/(m*n)),同时降低噪声、光照等对匹配结果的影响,提高相似度得分;在非顶层利用优化后的NCC公式结合积分图像进行相似度计算,计算量可减少约2/3。实验结果显示Edge-NCC算法在图像存在光照变化、噪声以及模糊的鲁棒性上相较于原有的NCC算法有了一定的改善,在这些不同条件下的匹配成功率能够达到约95%,验证了算法的有效性。在视觉点胶机的工件定位应用过程中,使用该算法能够有效地增加定位的准确度和降低企业的生产成本。
The traditional normalized cross-correlation(NCC) algorithm is based on grayscale implementation,which has the problems of large computation,ignoring the image feature information,and not being able to match the rotated image,etc.Therefore,an Edge-NCC template matching algorithm based on the OpenCV implementation that combines the image features and the grayscale information is investigated.The algorithm uses edge points for similarity calculation in the top layer of the image pyramid,and the computation amount can be reduced by(1-z/(m*n)),and at the same time,it reduces the influence of noise,illumination,etc.on the matching results,and improves the similarity scores;in the non-top layer,it utilizes the optimized NCC formula combined with the integral image for similarity calculation,and it can reduce the computation amount by about 2/3.The experimental results show that the Edge-NCC algorithm has a certain improvement compared with the original NCC algorithm in the robustness of the image in the presence of illumination changes,noise and blurring,and the matching success rate under these different conditions can reach about 95%,which verifies the effectiveness of the algorithm.The use of this algorithm can effectively increase the accuracy of positioning and reduce the production cost of enterprises in the process of workpiece positioning application of vision dispensing machine.
作者
田钊
张春雷
张敬宇
洪国
TIAN Zhao;ZHANG Chunlei;ZHANG Jingyu;HONG Guo(School of Mechanical Engineering,Sichuan University,Chengdu 610065,China)
出处
《组合机床与自动化加工技术》
北大核心
2025年第2期158-162,共5页
Modular Machine Tool & Automatic Manufacturing Technique
