【摘要】 针对管道弯曲度测量系统中照度不均匀的平面网格图像,提出了一种利用变分的自适应阈值分割方法。综合考虑图像的灰度分布和梯度信息,以图像的边缘点作为插值点,构造一个自适应的阈值曲面。在极小化阈值曲面能量函数时,使用变分方法化简Poisson方程,并利用逐次超松弛迭代法求方程的收敛解,从而产生稳定、光滑的阈值曲面,实现对照度不均匀的平面网格图像进行有效分割。为了验证该方法的有效性,进行了图像分割实验,选择超松弛参数为1.5,变分参数为26。并且与常用的最大类间方差法、迭代法和一维最大熵法进行了比较。实验结果表明,对于照度不均匀的平面网格图像,该分割算法的分割效果明显优于其他三种算法,它能准确地分割出平面网格图案。更多还原

【Abstract】 In view of a plane gridding image with nonuniform illumination in a camber measuring system, we propose an adaptive threshold method by the variational method. On the basis of integrating gray-level distribution and gradient information, we use boundary points as known data to interpolate the adaptive threshold surface of the image. In order to acquire the minimum of the energy function, a successive overrelaxation iteration method by the variational method is used to solve a Poisson equation. As a result, we can get a convergent solution, which keeps the threshold surface robust and smooth. In order to validate the proposed method, a test is implemented as an overrelaxation parameter is 1.5 and a variational parameter is 26, and the method is compared with the Otsu’s method, the iteration method and the one-dimensional maximum entropy method. Experimental results show that the proposed method has the virtue of correct segmentation for the plane gridding image with nonuniform illumination.更多还原