中文摘要：

针对传统灰度级-平均灰度级二维直方图区域划分存在的不足,提出了由灰度级-梯度构成改进的二维直方图区域划分方法,并给出了基于改进二维直方图的Otsu阈值法及其快速递推算法。实验结果表明：提出的改进二维Otsu法与传统二维Otsu法相比,抗噪性能好,分割所得的区域内部具有较好的一致性,获得的目标边缘更加清晰,同时减少了运行时间。

英文摘要：

According to the obvious deficiencies in the regional division of the traditional gray level-average gray level two-dimensional histogram, an improved regional division method of the gray level-gradient two-dimensional histogram is proposed and meanwhile an Otsu＇s threshold segmentation method and its fast recursive algorithms based on improved two-dimensional histogram are given. The experimental results show that this method achieves better anti-noise ability and the regions of segmentation bear a superior consistency, accurate borders and the computing time of the improved Otsu＇s method is less than the traditional one.