摘要
衬砌裂缝图像的分割和信息提取是裂缝自动检测系统中的重点。在采集图片大量试验的基础上对裂缝信息采集流程进行优化。其中,针对衬砌图像分割,提出结合prewitt算子和Otsu阈值的梯度类间阈值法,并用数学形态学对图片进行进一步优化,从而得到裂缝区域完整而杂质区域较少的二值图像;针对裂缝特征提取,通过对试验图片集合特征的分析确定描绘子为面积、周长、集中度、外接矩形长宽比、外接矩形占有率以及平均宽度。应用实例表明,梯度类间阈值法能够更好地实现对图像的分割,较常用的Otsu法更能适应复杂背景的裂缝图片。通过对优化图像的不同区域进行描绘特征提取统计,可形成具有一定代表性的图片库,为实现裂缝自动识别奠定基础。
The segmentation and characteristic extraction of lining crack image are essential for the automatic crack detection. The digital image processing analysis was optimized and some algorithms were improved based on the large number of collected images from tests. For the image segmentation,a new algorithm of binarization combining Prewitt operator with Otsu threshold was proposed. The binary image processed by the new algorithm,combined with the mathematical morphology resulted in the one with more complete fracture zones and less interference regions. For characteristic extraction, area, perimeter, compactness, rectangle degree, Feret's occupying rate and the mean width were defined as the morphological characteristics. The results of application showed that the integrated algorithm achieved the better image segmentation and was more suitable for the crack image with complex background. The characteristics extracted from the different regions in the image formed a representative information base and laid a foundation for automatic identification of cracks.
出处
《岩石力学与工程学报》
EI
CAS
CSCD
北大核心
2015年第6期1211-1217,共7页
Chinese Journal of Rock Mechanics and Engineering
基金
国家自然科学基金资助项目(51278423
51478395)
关键词
隧道工程
PREWITT算子
Otsu阈值
二值图像
特征提取
tunnelling engineering
Prewitt operator
Otsu threshold
image binarization
characteristic extraction
