基于直线拟合的高温板材边缘检测方法

提出了一种基于最小二乘原理的垂直拟合边缘检测方法。该方法针对高温板材热膨胀率测量过程中异常边缘点随机出现的特点,利用最小二乘法同时拟合两个相互垂直方向上的边缘直线,并进行多次重复拟合使边缘直线逐渐接近真实值。实验表明,该方法能在异常点存在的情况下准确、快速地拟合出高温钢板图像的边缘直线。

Edge Detection Method of Plates under High Temperature Based on Linear Fitting

An edge detection method of perpendicular fitting based on least square theory is present in this paper. According to the characteristics of the abnormal edge points’ random appearances, least square method is employed to fit two edge lines in the mutual perpendicular directions simultaneously, and the edge line gradually approaches true value through iterative fitting. Experiment result shows that edge of plate’s image under high temperature can be fitted accurately and fast in the condition that abnormal points still exist by the method proposed.