基于分数阶微分的消除局部强反射的孔径测量方法

Measurement of aperture based on fractional derivative for eliminating local strong reflection

针对视觉测量硬盘圆孔直径易受到圆孔周围高反光面强反射的影响而导致测量精度不高的问题,提出了一种基于分数阶微分的图像去噪声的处理方法。通过分数阶微分算法对相机采集的带有强反射、高反光的硬盘圆孔孔径图像进行处理,消除圆孔周围强反射表面等不相关信息对圆孔边缘提取带来的影响,通过实验分别与Prewitt、Soble、Laplacian算子进行比较,证明分数阶微分可有效减少所要分析的图像信息量,更好地增强圆孔边缘轮廓信息,可达到更好的视觉效果。对分数阶微分算法处理后的圆孔图像进行Canny边缘检测,提取出有效的圆孔边缘,利用最小二乘法对孔径边缘进行直径测量。实验结果表明:该文算法与其他算法相比,在保证精确提取孔径边缘信息的基础上,对圆孔周围的强反射面进行了抑制,误差控制在0.05mm左右,保证了测量精度。

In order to solve the problem that the hole diameter of the hard disk is easy to be affected by the local strong reflection and the high reflective surface,a method for measuring the diameter of circular parts to removal the noise based on the fractional differential was presented. The fractional-order differential algorithm was used to process the images of the hard disk aperture images with strong reflection and high reflection, so as to eliminate the influence of irrelevant information in strong reflection surface around circular hole on the extraction of the edge of circular hole. Experiments were carried out to compare with the Prewitt, Soble and Laplacian operators. It is proved that the fractional differential can ef-fectively reduce the amount of image information to be analyzed and achieve a better visual effect. Moreover the Canny edge detection was performed for the circular hole image after processed by fractionalorder differential algorithm to extract the available circular hole edge. Finally, the least square method was used to measure the hole size. Experimental results show that, compared with other algorithms, in the case of accurately preserving the edge information of the round hole and effectively suppressing the surrounding noise ,the error can be controlled around 0. 05 mm, and the the measurement accuracy is guaranteed.