The invention discloses a tunnel face geology multi-dimensional digitized
record recognition method. The method comprises the steps that a target object is determined, installation positions of a video shooting device, an
infrared thermometer and a
laser ranger finder are determined, the target object is divided into four parts which are imaged respectively, images are preprocessed, the preprocessed images are subjected to structural
information extraction and analysis, and structural surface features and rock
mass texture features are obtained, a three-dimensional
geological structure model is reestablished, geology is predicted, fracture surface and rock stratum joint images of adjacent faces are obtained from face images in the three-dimensional
geological structure model, every two adjacent rock stratum joint boundaries share the same structural surface, a virtual section
cutting plane at any angle crosses with the boundary of the corresponding structural surface, and a geology longitudinal plan is obtained. The three instruments are adopted to quickly, efficiently and comprehensively collect the conditions of the faces, the collecting mode that the images of partitions are collected respectively and then synthesized into images is adopted for the tunnel face images, the
image quality of the collected images can be improved,
image processing in a later stage is easy, and the accuracy of the images is improved.