沈阳工业大学学报2018，Vol.40Issue(2)：235-240,6.DOI:10.7688/j.issn.1000-1646.2018.02.21
基于特征分布的建筑墙体裂缝图像识别方法
Image recognition method for building wall cracks based on feature distribution
摘要
Abstract
Aiming at the problem that the current wall crack identification based on artificial intelligence is mainly composed of image recognition,which is easily affected by the uneven distribution of crack feature, and the identification accuracy is not high, a building wall crack image recognition method based on both the feature distribution and Gaussian mixture model was proposed. The corner solution processing for the wall image was carried out with the Harris corner detection algorithm, and the building wall image was preprocessed. Through selecting the mask smoothing method,the enhancement processing for the wall body image was performed. Through combining the feature distribution and Gaussian mixture model, the high-precision recognition of building wall crack image was realized. The results show that the proposed method has higher recognition precision and shorter identification period, and the pretreatment effect is obviously enhanced.关键词
墙体裂缝/图像识别/增强处理/局部直方图均衡法/高斯混合/Harris角点检测算法Key words
wall crack/image recognition/enhanced processing/local histogram equalization method/Gauss mixture/Harris corner detection algorithm分类
信息技术与安全科学引用本文复制引用
曹艳玲,袁义宏..基于特征分布的建筑墙体裂缝图像识别方法[J].沈阳工业大学学报,2018,40(2):235-240,6.基金项目
广东省教育厅本科高校教学质量与教学改革工程立项建设基金资助项目(粤教高函〔2015〕133号) （粤教高函〔2015〕133号）
华南理工大学中央高校基本科研业务费资助项目(2015BS09). （2015BS09）
