摘要： 提出了一种利用一组几何特征参数对图像中的圆形进行检测的方法。对提取出的边缘进行多边形拟合，并对拟合后的多边形进行归一化处理，提取平移、旋转和尺度变换不变性的几何特征，计算相应的几何特征参数，将几何特征参数满足一定条件的形状识别为圆形。描述了该方法的具体步骤，并与随机Hough变换进行了对比实验。对合成图像和真实图像进行的实验结果表明，该方法具有较高的效率和实用性。

关键词: 圆检测, 几何特征参数, 特征点, 图像分割, 边缘提取

Abstract: A circle detection method using a group of geometric feature parameters is proposed. Polygon fitting to the extracted edges is implemented and the fitted polygons are normalized to obtain invariant geometric features to translation, rotation, and scale change. The parameters corresponding to the geometric features are calculated. Those shapes with the geometric feature parameters satisfying certain conditions are recognized as circles. The algorithm is described, and comparison experiments with the randomized hough transformation are made. The experimental results on synthetic and real images show that the proposed method is efficient and practical.

Key words: Circle detection, Geometric feature parameter, Feature point, Image segmentation, Edge extraction