基于Zernike矩的亚像素边缘定位算法

亚像素软件处理技术可以在一定程度上补偿图像测量系统由硬件限制引起的边缘定位精度,提出了一种基于Zernike矩的亚像素边缘定位算法。该算法利用Zernike多项式的正交性与完备性以及复数矩幅度的旋转不变性计算出边缘的亚像素位置。计算出了像素为正方形与矩形(长宽比为4∶3)时的矩模板,使其应用范围进一步扩大。分析了由于矩模板产生的原理偏差,并提出了修正公式,使定位精度得到提高。进一步推导出了实际边缘模型的定位公式,导出了其存在的偏差,并进行了补偿。最终的实验结果表明,补偿后的算法对直线的定位精度达0.05像素,对曲线的定位精度可达0.07像素。

Algorithm for edge subpixel location based on Zernike moment

Subpixel algorithm is an effective method to improve the accuracy of edge location under assured hardware condition. A novel algorithm for edge subpixel location based on Zernike moment was proposed. Zernike polynomials were selected in this algorithm based on its orthogonality and completeness, and the rotation invariance of complex moment amplitude. The moment masks of square and rectangle (the aspect ratio is 4∶3) pixel was performed for the extending of its application. The principle deviation caused by moment masks was discussed. The modified formula was given for the improvement of the edge location accuracy. The subpixel location formula for practical edge was deduced. The location error was analyzed and compensated. The experimental result shows that the location accuracy using the Zernike algorithm after compensated is up to 0.05 pixels and 0.07 pixels for straight-line and curve respectively.