基于最小二乘法椭圆拟合的改进型快速算法

Improved Fast Algorithm for Elliptic Fitting Based on Least Square Method

采用最小二乘法6样本点重复多次随机拟合椭圆,选取最高匹配度拟合参数能够很好地剔除误差较大样本点带来的影响,但是在样本点较多的情况下需要重复计算匹配度的次数较多,极大增加了算法的计算时间。考虑这种情况在上述算法基础上增加样本点分区方法,减少重复拟合的次数,减少算法收敛时间,并通过Matlab仿真验证算法的可行性。

Using the least square method 6 sample points repeated several times random fitting ellipse,choosing the highest matching parameters can be good to eliminate the error of larger sample points brought by the impact.However,in the case of more sample points,it is necessary to repeatedly calculate the matching degree more times,which greatly increases the computing time of the algorithm.Considering this situation,the sample point partition method is added on the basis of the above algorithm of reduce the number of repeated fitting and the convergence time of the algorithm,the feasibility of the algorithm is verified by Matlab simula⁃tion.