Multi Point Visual Monitoring of Deformation in Deep Foundation Pits During the Construction of Comprehensive Buildings
A multi-point visual monitoring method for deformation of deep foundation pits in comprehensive building construction is proposed with the aim of ensuring construction safety.Taking a deep foundation pit project of a comprehensive building construction project to be tested as the research target,measuring points are set up around it and image sensors are deployed to collect relevant images of the top of the deep foundation pit support pile,surrounding roads,pipelines,etc.The three-dimensional deep foundation pit image is reconstructed using the bounding box method.The rough K-means clustering algorithm is used to calculate the range and center of the target points.The deformation displacement value of the deep foundation pit is obtained by comparing the difference between the original image and the monitoring image's landmark points,Construct monitoring data samples using the displacement changes of landmark points in the obtained deep foundation pit images,and output deformation prediction results of deep foundation pits through training an RBF neural network.The experimental results show that this method can effectively monitor the deformation of deep foundation pits;The maximum deformation of the road settlement around the deep foundation pit is about 15.12mm,which has not reached the alarm value of 40mm;The reinforcement structure of the enclosure can effectively constrain the deformation of deep foundation pits.
NETL
NSTL
万方数据
