中文摘要：

针对现有区间值模糊集的模糊熵度量的不足,提出该类模糊集的一种新的熵刻画,并将其应用到图像分割中.用区间值模糊集表示图像,依据最优阈值选取准则确定图像分割的初始阈值.结合Canny边缘检测结果研究一种针对水下图像的动态阈值分割算法,实现水下图像的目标检测.仿真实验结果表明该算法具有很好的图像分割和目标检测效果.

英文摘要：

In view of shortcomings with existing versions,a new characterization for interval-valued fuzzy entropy sets was proposed and applied to image segmentation.An image was represented by an interval-valued fuzzy set,and an initial threshold for image segmentation was determined by the principle of optimal entropy.In combining the initial threshold with the result of classical Canny edge detection,a novel dynamic thresholding technique for image segmentation was proposed to meet the aim of underwater image object detection in complex circumstances.Simulated examples show the rationality and practicality of the new technique for image segmentation and object detection.