基于遗传算法和简化PCNN的裂缝检测方法

发布时间：2018-06-07 16:04

本文选题：裂缝检测 + 简化脉冲耦合神经网络　；参考：《计算机应用研究》2017年06期

【摘要】：为能够在复杂背景下检测裂缝、分析裂缝图像特征,由脉冲耦合神经网络(pulse coupled neural networks,PCNN)的运行特征和神经元的状态变化分析简化PCNN模型,将简化PCNN模型用于裂缝图像的目标检测。针对PCNN无法确定裂缝图像的最优检测以及脉冲门限具有非线性因子的问题,提出了一种基于遗传算法(genetic algorithm,GA)和简化PCNN的裂缝图像检测方法。该方法采用最小误差准则作为遗传算法的适应度函数,并且根据遗传算法具有全局最优解的特点确定简化PCNN中各因子的值,实现了简化PCNN的裂缝图像自动分割。将该方法与不同的分割方法对实际裂缝图像的处理结果进行比较,通过区域对比度、准确率和召回率等客观指标进行定量分析,表明了该方法对裂缝图像检测的有效性与通用性。
[Abstract]:In order to detect cracks and analyze the features of crack images in complex background, the simplified PCNN model is simplified by analyzing the running characteristics of pulse coupled neural networks and the state changes of neurons, and the simplified PCNN model is used to detect the target of crack images. Aiming at the problem that PCNN can not determine the optimal detection of crack image and the pulse threshold has nonlinear factor, a crack image detection method based on genetic algorithm and simplified PCNN is proposed. In this method, the minimum error criterion is used as the fitness function of the genetic algorithm, and the value of the factors in the simplified PCNN is determined according to the characteristics of the global optimal solution of the genetic algorithm. The crack image segmentation of the simplified PCNN is realized automatically. The method is compared with different segmentation methods to deal with the actual crack image, and quantitative analysis is carried out through objective indexes such as regional contrast, accuracy rate and recall rate, etc. The effectiveness and generality of this method for crack image detection are demonstrated.
【作者单位】：郑州大学信息工程学院;
【基金】：国家科技支撑计划资助项目(2014BAH09F00) 国家自然科学基金资助项目(61379079)
【分类号】：TP183;TP391.41

【相似文献】