The invention discloses a tunnel back break intelligent detection method based on a three-dimensional
laser scanning technology, and relates to the field of
tunnel engineering construction. The method comprises the following steps: inputting a tunnel three-dimensional
laser scanning
point cloud data set collected on site, automatically carrying out fine-grained
processing and training learning by an
algorithm, automatically extracting data features, obtaining classification
model parameters, taking the classification
model parameters as fixed values of a
machine learning
algorithm model, segmenting each three-dimensional
point cloud model into
point cloud pieces, comparing point cloud
label information, and carrying out
machine learning. And
field point clouds with the same characteristics are merged, manifold is introduced as a constraint condition, and a tunnel section over-excavation and under-excavation three-dimensional
laser scanning model is reconstructed by means of a curved
surface construction method. According to the method, massive point
cloud data is intelligently processed, so that the constructed tunnel curved surface integrally meets Euclidean spatial characteristics, the local characteristics of the tunnel section model are guaranteed, the real spatial form of the tunnel is more accurately reflected, accurate judgment of over-excavation and under-excavation in the tunnel section construction process is realized, and a powerful
technical support is provided for safe construction of the tunnel.