基于超模糊熵ULPCNN二值图像分割算法研究与实现  

Research and Realization Based on Fuzzy Entropy Algorithm ULPCNN Binary Image Segmentation

作　　者：唐土生[1] 陈绚青[2] 

机构地区：[1]常州工学院,常州213002 [2]江苏理工学院,常州213002

出　　处：《激光杂志》2016年第1期113-116,共4页Laser Journal

基　　金：常州市科技支撑项目(CQ2007005)

摘　　要：图像的精确分割能够加快后期图像的处理,在实际图像处理中具有广泛的应用。脉冲耦合神经网络虽在处理图像分割、图像平滑度中具有众多优点,但存在神经元点火状况相对复杂,对全局图像阀值不能清楚的展现等缺点。针对以上缺点本文提出了一种Unit-Linking PCNN(ULPCNN)抑制捕获模型,在模型中引入了阀值函数该函数呈指数上升;在单纯的PCNN图像分割中引入超模糊熵得到了基于超模糊熵ULPCNN二值图像自动分割算法,实现对图像数据的模糊性和不精确性进行有效处理,对图像进行自动分割。并与最大香农熵图像分割法、最小交叉熵图像分割法以及最小模糊的ULPCNN图像分割法等进行了比较,实验仿真说明本文提出的分割图像算法更加可靠有效。Accurate plications in the actual image segmentation can speed the processing of images in the late, having a wide range of apimage processing. Pulse coupled neural network in image processing segmentation, image smoothness has many advantages, but there are some disadvantoges likeneurons ignition condition is relatively complex, on the global image threshold cannot clearly show the disadvantages. For the above shortcomings, this paper proposed a unit linking PCNN（ulpcnn） inhibit the capture model. The threshold function which rises exponentially is introduced in- to the model. The superfuzzy entropy is introcluced into the simple PCNN image segmentation, the supermodel fuzzy en- tropy ULPCNN binary image automatic segmentation algorithmis obtcined, effective treatment for achieve the ambiguity and imprecision of image data , the image automatic segmentation. And camparing with a maximum Shannon entropy image segmentation method and minimum cross-entropy image segmentation methods, minimum fuzzy ULPCNN image segmentation method, simulations show the segmentation algorithm is more reliable and efficient

关 键 词：图像分割 脉冲耦合神经网络 阀值函数 超模糊熵 

分 类 号：TN911.73[电子电信—通信与信息系统]