首页|基于RGBD图像的预制构件钢筋间距高效检测方法

基于RGBD图像的预制构件钢筋间距高效检测方法

扫码查看
为保证预制构件的生产质量且在现场可顺利安装,通常需要在出厂前进行钢筋间距检测.现有的钢筋间距检测以人工检测为主,人力及时间成本高.为提高钢筋间距检测效率,提出 一种基于RGBD图像的预制构件钢筋间距高效检测方法.基于深度相机采集的彩色图像,利用语义分割神经网络实现钢筋像素高效分割;利用相机内部参数生成点云,根据钢筋像素精确分割钢筋点云并基于点云特征进行数据增强;使用点云处理算法,实现对预制构件钢筋间距的高效检测.在验证试验中,对一块包含272根钢筋的预制混凝土板单元的出筋间距进行了检测.结果表明,所提出的检测方法能够准确高效地完成预制构件的钢筋间距检测,具有显著的实用价值及经济效益.
Efficient detection method for rebar spacing of prefabricated components based on RGBD images
To ensure the production quality of prefabricated components and the smooth installation on site,it is usually necessary to detect the spacing of steel bars before leaving the factory.Current rebar spacing detection primarily rely on manual methods,which are time-consuming and labor-intensive.To improve the efficiency of rebar spacing detection,an efficient detection method of rebar spacing of prefabricated components based on RGBD image was proposed.Based on the color image collected by the depth camera,the semantic segmentation neural network was used to achieve efficient segmentation of rebar pixels.The point cloud was generated by using the internal parameters of the camera,and the point cloud of the rebar was accurately segmented according to the steel bar pixels and the data was enhanced based on the point cloud features.The point cloud processing algorithm was used to realize the efficient detection for rebar spacing of prefabricated components.In the verification test,the rebar spacing of a precast concrete slab unit containing 272 rebar was tested.The results show that the proposed detection method can accurately and efficiently complete the detection of rebar spacing of prefabricated components,and has significant practical value and economic benefits.

RGBD imagesemantic segmentationpoint cloud datarebar spacingefficient detectiondepth camera

李智鹏、李江、李东声、刘界鹏、陈奉民

展开 >

重庆大学山地城镇建设与新技术教育部重点实验室,重庆 400045

重庆大学土木工程学院,重庆 400045

中铁长江交通设计集团有限公司,重庆 401121

RGBD图像 语义分割 点云数据 钢筋间距 高效检测 深度相机

2024

建筑结构
中国建筑设计研究院 亚太建设科技信息研究院 中国土木工程学会

建筑结构

CSTPCD北大核心
影响因子:0.723
ISSN:1002-848X
年,卷(期):2024.54(24)