基于机器视觉的型钢梁施工过程变形监测技术研究

Deformation Monitoring Technology of Steel Beam Construction Based on Machine Vision

针对施工条件下的实际工程情况,提出一种颜色阈值匹配结合霍夫变换的轻量级机器视觉算法对型钢梁变形进行监测。首先,通过RGB转换HSV空间得到颜色阈值,之后进行霍夫变换识别,结合图像处理操作,可同时得到多个目标靶点的位置信息进行变形监测。其次,现场试验验证表明,机器视觉技术可在复杂背景下取得较好效果,最终实现在温度作用下型钢梁变形位移监测。最后,数据分析表明,温度对型钢梁的变形在安全范围内,不影响施工进度,且提出算法在精度和稳定性方面均可满足工程要求。

Aiming at the actual engineering situation under construction conditions,a lightweight machine vision algorithm based on color threshold matching and Hough transform is proposed to monitor the deformation of steel beams.Firstly,the color threshold is obtained by RGB conversion to HSV space,and then Hough transform is performed to identify.Combined with image processing operations,the position information of multiple targets can be obtained at the same time for deformation monitoring.Secondly,the field test verification shows that the machine vision technology can achieve good results under complex backgrounds,and finally realize the deformation and displacement monitoring of steel beams under the action of temperature.Finally,data analysis shows that the deformation of the steel beam due to temperature is within a safe range and does not affect the construction progress,and the proposed algorithm can meet the engineering requirements in terms of accuracy and stability.