水泥稳定碎石摊铺均匀性的图像检测技术研究

针对水泥稳定基层摊铺施工均匀性检测主要依靠人工目测判别,存在检测精度低和判断结果主观等问题,文章采用图像处理技术进行摊铺均匀性检测,建立基于深度学习的水泥稳定碎石颗粒图像分割方法,将图像中各档集料分离,结合四边静矩理论检测图像中的粗集料摊铺施工均匀性。检测结果表明:该方法能有效地分割水泥稳定碎石基层摊铺图像,依据图像数据客观地检测集料分布的均匀性状况,为水泥稳定基层摊铺施工均匀性评价提供依据。

In response to the fact that the uniformity detection of cement stabilized base paving construction mainly relies on manual visual inspection, which has problems such as low detection accuracy and subjective results, image processing technology is used for paving uniformity detection. A deep learning based image segmentation method for cement stabilized macadam particles is established, which separates the various grades of aggregates in the image and combines the quadrilateral static moment theory to detect the uniformity of coarse aggregate paving construction in the image. The detection results indicate that this method can effectively segment the image of cement stabilized macadam base paving, objectively detect the uniformity of aggregate distribution based on image data, and provide a basis for evaluating the uniformity of cement stabilized base paving construction.