基于数字图像处理技术的建筑预埋件无人化识别技术

Unsupervised Identification Technology of Building Embedded Parts Based on Digital Image Processing Technology

在建筑施工质量检查中,需对预埋件的位置及数量牢牢把控。为进一步提高对建筑预埋件的检测效率,利用数字图像处理技术,搭建并训练形成了建筑预埋件识别的卷积神经网络模型。基于此,通过配置相关硬件系统,开发了一款可移动式机器人,并内嵌预埋件识别模型,形成了建筑预埋件的无人化识别技术。最后,将上述无人化识别技术应用于实际工程中,实现了对建筑预埋件的有效检测及数量统计,验证了技术的有效性。

In the inspection of construction quality,it is important to confirm the position and quantity of embedded parts.To further improve the detection efficiency of embedded parts and realize unsupervised identification of building embedded parts,this paper adopted digital image processing technology,and built a convolution neural network model for building embedded parts recognition.Furthermore,a mobile robot was developed which was embedded in the convolution neural network model.Thereafter,the unsupervised identification technology of building embedded parts was formed.Finally,the unsupervised identification technology mentioned above was applied to the actual project which detected the embedded parts and acquired the quantity of them effectively.