摘要
准确评估大坝裂缝是保证大坝安全的关键,对此提出一种基于图像处理与轻度梯度提升树的混凝土坝表面裂缝逐像素点自动识别与分类模型,并以自制水工混凝土表面裂缝图像数据集为例进行分析。试验结果表明,所提方法可实现混凝土坝表面裂缝识别模型的快速建模,完成复杂背景下裂缝轮廓的准确提取。该方法可有效解决数据匮乏时大坝裂缝识别模型建模困难的问题,为大坝安全巡检工作开展提供支撑。
Accurate assessment of dam cracks is the key to ensuring dam safety.In this study,an automatic pixel-by-pixel classification and detection model of concrete dam surface cracks is proposed based on image processing and light gradient lifting tree.Then a self-made dataset containing hydraulic concrete surface crack images is used as the case study.The experimental test results show that the proposed method can realize the rapid modeling of the surface crack recognition model of concrete dams,and complete the accurate extraction of crack contours under complex backgrounds.This method can effectively solve the problem of difficulty in modeling the dam crack recognition model when data is scarce,and provide support for the development of dam safety inspection work.
作者
高治鑫
包腾飞
李扬涛
GAO Zhi-xin;BAO Teng-fei;LI Yang tao(College of Water Conservancy and Hydropower Engineering,HohaiUniversity,Nanjing 210098,China;State Key Laboratory of Hydrology Water Resources and Hydraulic Engineering,HohaiUniversity,Nanjing 210098,China;College of Hydraulic&Environmental Engineering,China Three Gorges University,Yichang 443002,China)
出处
《水电能源科学》
北大核心
2022年第4期95-98,共4页
Water Resources and Power
基金
国家重点研发计划(2018YFC1508603)
国家自然科学基金项目(51739003)
江苏省研究生科研与实践创新计划项目(KYCX21_0515)。
关键词
裂缝识别
机器学习
语义分割
轻度梯度提升树
crack identification
machine learning
semantic segmentation
LightGBM
