摘要
研究了一种基于二次曲线的椭圆估计方法,该方法的主要优点是在进行椭圆中心点检测的时候,不用检测图像上单个椭圆的边界,而只计算图像上的梯度信息。利用椭圆的梯度信息计算椭圆上各点的切线方程,根据切线方程和椭圆的二次曲线性质解算出椭圆曲线的各个系数,根据这些系数就可以求得椭圆的中心点及其各个参数。并且通过引进形态学的图像处理方法,对梯度的计算进行改进,使最终结果达到很好的精度。从最后三维重建的结果可以看出,整体误差在0.069mm,标准差0.018mm,检测的精度高,能满足一般测量的要求。
The method for ellipse estimation based on conic is researched in this paper. The main advantage of this method is that only the gradient information of the image need to be calculated instead of the boundary of individual ellipse when the center of ellipse is detected. By using the gradient information of ellipse , the tangential equations of different points on ellipse are obtained. According to tangential equation and the features of ellipse conic , the coefficients of ellipse curvilinear equation can be calculated, and then the center and the parameters of ellipse can be got.. The gradient calculation can be improved through introducing the image processing method of morphology, and the accuracy of final results can be increased. The results of 3d reconstruction show that the overall error is 0.069 mm, and the standard deviation is 0.018mm. The detection precision is high and can meet the requirements of general measurements.
出处
《传感器世界》
2011年第6期10-14,共5页
Sensor World
基金
四川省科技厅国际合作资助项目(项目编号:2009HH0023)
关键词
二次曲线
梯度
椭圆检测
形态学
conic
gradient
detection of ellipse
morphological
