CAJ | 학술논문

　　针对当桥梁体建筑开裂程度细微或者裂痕的特征发生混淆时，以此建立的高斯模型会发展特征混合，很难准确区分待识别特征，为了解决这一问题，提出一种人工免疫算法优化SVC的计算机裂痕识别方法，通过对裂痕特征进行预处理，建立支持向量机的模型对混叠的细微特征进行分类，运用人工免疫算法对分类的效果进行优化处理，解决由于灰度特征发生错误，桥梁桥体裂痕识别准确度不高的问题。实验证明，这种裂痕识别算法实现简单，能够克服裂痕特征过于细微带来的缺陷，得到了较好的识别效果。
　　침대당교량체건축개렬정도세미혹자렬흔적특정발생혼효시，이차건립적고사모형회발전특정혼합，흔난준학구분대식별특정，위료해결저일문제，제출일충인공면역산법우화SVC적계산궤렬흔식별방법，통과대렬흔특정진행예처리，건립지지향량궤적모형대혼첩적세미특정진행분류，운용인공면역산법대분류적효과진행우화처리，해결유우회도특정발생착오，교량교체렬흔식별준학도불고적문제。실험증명，저충렬흔식별산법실현간단，능구극복렬흔특정과우세미대래적결함，득도료교호적식별효과。
For when the bridge wall building cracking degree fine or crack characteristics occur when confusion, this pa-per builds up the gaussian model will be mixed development characteristics, it is difficult to accurately distinguish to i-dentification characteristics, in order to solve this problem, the paper proposes a kind of artificial immune algorithm to optimize the SVC computer identification of crack identification method, through the crack characteristics of pretreatment, through the establishment of support vector machine (SVM) model of aliasing subtle features of classification, the use of the artificial immune algorithm for classification effect optimized, solve the grayscale characteristics mistakes, Bridges wall crack identification accuracy is not high question. Experiments show that the crack identification algorithm is simple, can overcome crack characteristics is too small to bring defects and obtain the better identification effect.