基于Zernike正交矩的亚像素图像线宽测量算法

Measurement algorithm for sub-pixel line width in images based on Zernike moment

针对图像内的细窄线宽,提出了一种基于Zernike正交矩的亚像素图像线宽检测算法。该算法具有明确的几何模型,通过计算图像的2阶和4阶Zernike正交矩,推导出了亚像素线宽表达式。根据数字图像的离散性,给出了计算正交矩所需的模板系数,并分析了由离散性造成的原理误差。将所提出的亚像素线宽检测技术应用于安瓿内异物粒径的标定实验,结果表明:该算法可有效地测量弱小目标在图像中的亚像素线宽值,从而得到了异物粒径大小与亚像素线宽之间的标定曲线。

A measurement algorithm based on Zernike moment for measuring thin line in images is presented. The algorithm has a definite geometric model. The expression of sub-pixel line width is derived by calculating second-order and fourth-order orthogonal Zernike moments of an image. According to the discreteness of digital images, a template coefficient is presented for calculating orthogonal moments and the principle error caused by the discreteness is analyzed. The proposed sub-pixel line width detection algorithm has been used in the calibration experiment for particle diameters in ampoule. The result shows that the sub-pixel line width of small particles in images can be efficiently measured by using this algorithm and the curve of sub-pixel line width vs particle diameter is obtained.