Abstract: The key point of traffic sign recognition is whether the used image detection algorithm has fast and high reliability or not.Based on color segmentation，a new circular traffic sign detection algorithm is proposed by using curve fitting.Color segmentation is used to remove most of the background subjects in the image.Then edge detection along with nonlinear least square method is used to detect the exact circular traffic sign based on its circular edge.The effectiveness of the algorithm is demonstrated preliminarily by computer simulation.

Key words: circle detection, color segmentation, curve fitting, nonlinear least square

摘要： 快速、可靠的交通标志检测是对其进行准确识别的前提，以颜色分割为基础，提出了一种基于曲线拟合的圆形交通标志检测算法。首先利用交通标志的颜色特征预分割出潜在的交通标志区域，然后针对圆形交通标志轮廓具有圆形这一关键特征，通过边缘检测并采用非线性最小二乘技术准确的确定出图像中的圆形交通标志区域。实验结果表明了算法的有效性。

关键词: 圆检测, 颜色分割, 曲线拟合, 非线性最小二乘