椭圆对象参数的快速提取  

Fast Method for Parameters Extraction of Elliptic Objects

作　　者：徐则中[1] 游庆祥[1] 钱诚[1] XU Zezhong;YOU Qingxiang;QIAN Cheng(School of Computer Science and Information Engineering,Changzhou Institute of Technology,Changzhou 213032,China)

机构地区：[1]常州工学院计算机信息工程学院,江苏常州213032

出　　处：《江苏海洋大学学报（自然科学版）》2020年第4期1-7,共7页Journal of Jiangsu Ocean University:Natural Science Edition

基　　金：国家自然科学基金资助项目(61602063);江苏省文化创意协同创新中心项目(XYN1705)。

摘　　要：针对椭圆检测的实时性需求,提出了一个快速的、基于区域像素的椭圆参数提取算法。该算法选用直线投票方程,运用区域像素,向两个1D数组进行Hough投票。推导了投票值与投票距离的函数关系。在两个投票方向上分别拟合一个二次函数,计算投票距离均值、投票距离宽度和最大投票值。运用投票距离均值计算椭圆对象的中心,运用投票距离宽度和最大投票值计算椭圆长轴和短轴,运用区域像素的统计特征计算椭圆对象的主轴方向。分别用模拟图像数据和现实图像数据对提出的椭圆检测方法进行测试,实验结果表明,该方法能够快速地从图像中提取椭圆对象的参数。所提出的基于区域像素的椭圆参数提取算法不需要进行边缘检测,仅向两个1D数组进行投票,因此提高了运算效率。In order to improve the time efficiency of ellipse parameter detection,a fast method for ellipses detection is proposed based on region pixels.The Hough voting equation is selected,and region pixels are used to vote for two one-dimensional arrays.The functional relationship between the voting value and the voting distance is derived.For each voting angle,the voting mean,voting width and maximum voting value are obtained by fitting a quadratic function.The center of the ellipse object is calculated based on voting means,the major axis and minor axis of the ellipse object are calculated based on the voting widths and maximum voting values.The orientation of the ellipse object is calculated based on the statistical characteristics of region pixels.The proposed ellipse detection method is tested with simulated image data and real image data respectively.The experimental results show that the proposed method can quickly extract ellipse object parameters from the image.The ellipse parameter extraction algorithm based on region pixels does not need edge detection,and only votes to two 1D accumulators,so the computational efficiency is improved.

关 键 词：椭圆检测 HOUGH变换 边缘轮廓 区域像素 函数拟合 统计特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]