为快速有效评价临床使用内窥镜图像相对畸变的程度, 提出一种基于椭圆拟合的检测方法。对内窥镜采集到的标准板图像进行二值化处理和图像分割; 运用anny算子对图像进行边缘检测并提取坐标值, 通过最小二乘法拟合椭圆, 求得椭圆参数; 计算相对畸变率。对3支内窥镜进行了椭圆拟合和定位质心相对畸变率检测, 结果显示相对畸变率在30%左右, 一致性差在2%左右, 椭圆拟合相较而言稳定性更好。

In order to evaluate the degree of relative distortion rate of clinical endoscopic images, adetecting method based on ellipse fittingalgorithm is proposed. First, the standard plate image captured by endoscope is processed by binary processing and segmentation pretreatment, the edge and coordinate of the image are detected by using canny operator. Then the ellipse is fitted by the least square method, and the parameters of the ellipse are obtained. Finally, the relative distortion rateis calculated. The relative distortion rate of 3 sets of endoscopes used ellipse fittingalgorithmand locating the center of mass method aredetected, and the results show that the relative distortion rate is less than 30%, and the consistency is less than 2%. By contrast, the stability of ellipse fittingalgorithm performs better.