欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
现有的Arimoto熵阈值法未考虑图像目标和背景的类内灰度均匀性,为此提出基于蜂群优化和基于分解的二维Arimoto灰度熵阈值分割方法.定义Arimoto灰度熵,导出二维Arimoto灰度熵阈值法,分别利用基于蜂群优化和基于分解的方法求解最佳阈值.基于蜂群优化方法给出中间变量的快速递推公式,利用改进的人工蜂群（MABC）优化算法搜索最佳阈值,减少迭代时适应度函数中的冗余运算.基于分解方法将求解二维Arimoto灰度熵阈值法的运算转化到2个一维空间,进一步降低计算复杂度.实验结果表明：与近年来提出的3种同类方法相比,所提出方法的分割性能更优,分割后图像中目标完整、边缘纹理清晰,具有良好的抗噪性.同时,所提出的方法运行速度快,有望满足实际系统对分割的实时处理要求.
英文摘要：
The existing thresholding methods based on Arimoto entropy do not consider the uniformity of gray scale within object cluster and background cluster.A 2D Arimoto gray entropy thresholding method based on bee colony optimization or decomposition was proposed.Arimoto gray entropy was defined and a2 D Arimoto gray entropy thresholding method was derived.The method based on bee colony optimization and another method based on decomposition were adopted to find the optimal thresholds.Fast recursive formulae for the intermediate variables were given using the method based on bee colony optimization.A modified artificial bee colony（MABC）optimization algorithm was adopted to find the optimal threshold of the 2DArimoto gray entropy method.The redundant computations of fitness function in an iterative procedure could be avoided.Using the method based on decomposition,the computations of 2D Arimoto grayentropy thresholding method were converted into two one-dimensional spaces.The computational complexity was further reduced.The experimental results show that,compared with three similar methods proposed recently,the proposed methods have superior image segmentation performance and a better antinoise performance.In the segmented images,objects are completely kept,and the edges and textures are clear.Moreover,the proposed methods have high running speed and can meet the real-time processing requirement of segmentation in the actual system.
同期刊论文项目
同项目期刊论文
期刊信息
位置：