【摘要】 在解释和说明Li与Lee提出的一维交叉熵阈值分割方法的基础上,将其推广到二维灰度直方图上,提出了二维交叉熵图像分割算法.为了克服二维空间上运算复杂性高、运算量大的缺点,给出了二维交叉熵阈值法的快速递推公式.与二维Otsu法相比,二维交叉熵阈值法能够更好地适应目标和背景方差相差较大的情形,是一种有效的阈值分割方法.更多还原

【Abstract】 Based on the explanation and illustration of the one-dimensional cross-entropy thresholding segmentation method proposed by Li & Lee,a two-dimensional cross-entropy thresholding method was proposed.Considering the disadvantage of high complexity and large computation of two-dimensional method,a fast recursive formula was presented also.Compared with the traditional two-dimensional Otsu’s method,the new method can be better adapted to the cases that the variance between the object and the background is large.It is validated that the new method is an available thresholding selection method.更多还原