基于绝对关联度的图像目标边缘提取方法

Image edge extraction based on absolute correlation degree

针对传统边缘检测方法对光学经纬仪实际拍摄的含噪声图像边缘检测效果不理想的问题,将灰色系统理论引入图像边缘检测,通过绝对关联分析与边缘检测算子相结合,提出一种利用二阶微分算子作为参考序列,图像像素作为比较序列,通过计算灰色绝对关联度来检测图像边缘的方法。实验结果表明,该方法与现有方法相比,边缘检测效果更好,检测的边缘清晰、连续,对噪声图像边缘的检测精度较高,抗噪声能力强。

Since the traditional edge detection methods can not get an ideal edge detection effect of the noisy images taken by optical theodolites,the gray system theory is applied to the image edge detection,and a method to detect the image edge by calculating the gray absolute correlation degree is proposed,which combines the absolute correlation analysis and edge detection operator,and takes the second-order differential operator as the reference sequence and image pixels as comparison sequence. The experimental results show that,compared with the available methods,this method can suppress the noise interference more effectively,its detecting effect is better,and the clarity and continuity of the image edge detected by the method are more ideal.