摘要
脚手架应用过程中,纵横间距不一致,导致重合交点处安全间距的风险值较高,存在难以确认横杆间距和立杆间距的问题。针对这一问题,提出基于像素距离和VGG16网络的脚手架安全间距检测方法。应用VGG16网络进行脚手架骨架的提取,并通过引入SE(Squeeze-and-Excitation)注意力模块提高特征提取能力,将提取的骨架图根据双阈值自适应二值化方法进行处理,以减少干扰并保留关键信息,通过计算像素点的4邻域之和标记潜在脚手架交点,并删除重合交点,在处理后的脚手架交点集合中,计算相邻交点之间的欧式距离,分析重合交点处安全间距的风险,并通过标靶法将像素距离转换为实际距离,以判断脚手架的各间距是否符合安全标准,实现了脚手架安全间距检测。实验结果表明:该方法能够提取出清晰、连续的脚手架骨架图,检测得到的横杆间距和立杆间距与实际间距的差值均在±5mm以内。
During the application of scaffolding,the uneven vertical and horizontal spacing results in a high risk of safety distance at the intersection point,making it difficult to confirm the spacing between horizontal bars and vertical bars.To address this issue,a scaffolding safety distance detection method based on pixel distance and VGG16 network is proposed.The VGG16 network is applied to extract the skeleton of scaffolding,and the SE(Squeeze and Excitation)attention module is introduced to improve the feature extraction ability.The extracted skeleton map is processed using a dual threshold adaptive binarization method to reduce interference and preserve key information.Potential scaffolding intersection points are marked by calculating the sum of the four neighborhoods of pixel points,and overlapping intersection points are deleted.In the processed set of scaffolding intersection points,the Euclidean distance between adjacent intersection points is calculated,and the risk of safe spacing at overlapping intersection points is analyzed.The pixel distance is converted into actual distance through the target method to determine whether the spacing of scaffolding meets safety standards and achieve scaffolding safety spacing detection.The experimental results show that this method can extract clear and continuous scaffold skeleton diagrams,and the difference between the detected horizontal and vertical spacing and the actual spacing is within±5 mm.
作者
郭爽
GUO Shuang(China Railway 18th Bureau Group No.4 Engineering Co.,Ltd.,Tianjin 300350,China)
出处
《机械工程与自动化》
2026年第1期157-160,共4页
Mechanical Engineering & Automation
