基于膨胀/腐蚀运算的神经网络图像预处理方法及其应用研究

Cellular Neural Network Applicating Manner in Pre-Processing Image

数学形态学在数字图像处理领域中的应用越来越广泛,各种形态结构和算法不断涌现.数学形态学以集合运算为基础,其基本思想是用具有一定形态的结构元去度量图像中的形态以解决理解问题.该文利用细胞神经网络(CNN),运用数学形态滤波适时、并行完成各种数学形态运算.文中给出了有关CNN的全局和局部稳定状态的定理,证明CNN在一定的条件下可以通过动态过程的稳定达到数学形态滤波的结果.将其结果运用在煤矿井下煤仓图像的预处理当中,取得了满意的结果.

Mathematical morphology is widely applicated in digital image processing.Variary morphology construction and algorithm being developed are used in deference digital image processing. The base idea of mathematical morphology is to use construction element measure image morphology for solving understand problem.The paper presents advanced Cellular neural Network that forms MMCNN equation to be suit for mathematical morphology filter.It gives the theories of MMCNN dynamic extent and stable state.It is evidenced that arrived mathematical morphology filter through steady of dynamic precess in definite condition.