PCNN模型具有相似群神经元同步发放脉冲的特性,适合于图像分割.对彩色图像的亮度分量进行对数变换,使其更符合人眼的视觉特性;在PCNN进行彩色图像R、G、B三分量分割的过程中,利用遗传算法进行神经元关键参数的选择,利用偏态指标进行迭代控制;在Unit-Linking PCNN模型中实现R、G、B三分量分割图的边缘检测,利用加权合并策略得到最终的边缘检测结果.仿真结果表明,该方法得到的结果体现了图像中更多的轮廓细节,具有很好的自适应性.%PCNN model has similar group neurons synchronization release pulse characteristic,appropriate for the image segmentation.Logarithmic transformation is done to the brightness component of color images to make color image much more fit for human's visual characteristic; during the process of PCNN segmenting R、 G、B three-components of color image, key parameters of the neuron are chosen by using genetic algorithm and iteration control is done through bias-normal distribution index;in the Unit-Linking PCNN model edge detection of R、G、B three-components segmentation image is achieved,and the final result of edge detection is acquired by using weighted merging strategy.The simulation results show that,the result can reflect more image outline and details,and the method has good adaptability.
展开▼
