基于细胞神经网络的楼面裂痕诊断研究  

Diagnosis of Floor Cracks Based on Cellular Neural Network

作　　者：张福新[1] 李国东[1] 

机构地区：[1]新疆财经大学应用数学学院,新疆乌鲁木齐830012

出　　处：《甘肃科学学报》2014年第5期1-5,共5页Journal of Gansu Sciences

基　　金：国家教育部人文社会科学基金(13YJAZH040);新疆维吾尔自治区高校科研计划项目(XJEDU2013I26)

摘　　要：为了更好地监测楼面裂痕,提出了一种改进细胞神经网络楼面裂痕图像识别方法.该方法通过一定的图像处理,建立裂痕网络和细节网络,同时增加了裂痕相似网络模型,避免了仅对裂痕特征提取信息不能准确识别裂痕的问题.实验证明,改进的裂痕识别算法实现简单,识别楼面裂痕准确率高,达到了实时识别技术的要求.An improved method for floor rift image identification based on cellular neural networks was developed to monitor floor cracks.Through image processing,rift networks and detail networks were created,and the rift similarity network model was established which avoids the problem that information concerning cracks is obtained but cracks are not accurately identified.Experiments proved that the improved method,more effective in identification computing and more accurate in detecting floor cracks,reaches the technical standard of real-time identification.

关 键 词：裂痕诊断 楼面裂痕 细胞神经网络 

分 类 号：TU37[建筑科学—结构工程]