Research on intelligent inspection method for buildings based on 3D vision
Illegal reconstruction of self-built buildings in urban and rural areas occurs from time to time,which is a hidden danger that cannot be ignored.Traditional building inspection mainly relies on manual methods,which have problems such as low efficiency and high subjectivity,thus affecting the reliability of inspection results.This paper proposes an intelligent method for building inspection based on 3D vision.Based on the technology of fusion SLAM,the point cloud data are collected in real time.Multiple inspection data are registered,and the radius search method based on kd-tree is used to identify the increase or decrease parts of the point cloud and find out the reconstructed parts.The reconstructed parts are segmented by region growing algorithm and the OBB bounding boxes of them are obtained,and simple component classification is performed according to the geometry information of the bounding boxes.Furthermore,the surface area change rate parameter is defined considering the importance of the components,which can preliminarily evaluate the safety of the reconstructed structure.Taking the inspection of a self-built building as an example,the results demonstrate that the method outlined above can efficiently and quickly obtain point cloud data of the building facade to be inspected.This enables effective identification of any altered parts of the building and allows for an assessment of the danger of the building based on the characteristics of the components,thus realizing the automation and intelligence of the inspection process.
building inspection3D visionpoint cloud registrationreconstruction identificationcomponent classificationrisk assessment
国家科技期刊平台
NSTL
万方数据
