基于支持向量机与BP神经网络的裂缝识别算法研究  

A study on the algorithms of crack recognition based on support vector machine and BP neural network

作　　者：孔燕 袁辉 柯敏勇[3] 范宇鑫 赵启林 潘大荣[4] Kong Yan;Yuan Hui;Ke Minyong;Fan Yuxin;Zhao Qilin;Pan Darong(Jiangsu Taizhou Bridge Co.,Ltd.,Taizhou 225300,China;Nanjing Tech University,Nanjing 210000,China;Nanjing Hydraulic Research Institute,Nanjing 210000,China;Nanjing Institute of Technology,Nanjing 210000,China)

机构地区：[1]江苏泰州大桥有限公司,江苏泰州225300 [2]南京工业大学,江苏南京210000 [3]南京水利科学研究院,江苏南京210000 [4]南京工程学院,江苏南京210000

出　　处：《山西建筑》2020年第8期129-131,共3页Shanxi Architecture

摘　　要：为了研究训练样本数量对裂缝识别算法效果的影响,在少样本和多样本的训练条件下建立了基于支持向量机和BP神经网络两种裂缝识别算法。两种算法的对比测试结果表明:在少样本情况下基于支持向量机的算法识别效果明显优越于基于BP神经网络的算法,适合工程使用初期、裂缝样本较少的情况;在多样本情况下两种算法的识别效果基本一致,均可以用于工程上。To study the effect of the number of training samples on the crack recognition algorithm,two crack recognition algorithms based on support vector machine and BP neural network are established with small-scale samples and large-scale samples.The comparative tests results showed that the algorithm based on support vector machine performed better than the algorithm based on the BP neural network in the case of small-scale training set,which is suitable for the case where there are few crack samples at the beginning of the project.The classification rate of the two algorithms is basically the same in the case of large-scale training set,which are both suitable for engineering applications.

关 键 词：裂缝识别 混凝土 支持向量机 BP神经网络 

分 类 号：TU312.3[建筑科学—结构工程]