Research And Implementation Of Tunnel Face Crack Extraction Algorithm Based On Digital Image

In the tunnel project,the tunnel safety accident has always been a concern in the construction process,and the stability and quality evaluation of the rock mass structure is the guarantee for the safety of the entire project design and planning.The face formed during the tunnel excavation contains a large amount of geological information,and rock fracture is the most important one.By extracting and analyzing cracks on the tunnel face,the accuracy of rock mass disaster prediction can be effectively improved.Only relying on human resources to collect these cracks is inefficiency,but the digital extraction method can greatly improve work efficiency and help geological personnel predict disaster accidents.By studying the extraction and analysis algorithms of fracture information on the digital image of the tunnel face,the entire tunnel project gains the most powerful guarantee.Rock structure information is the key to evaluate its stability and quality.Therefore,the extraction and analysis of tunnel face cracks is very important for practical engineering.In the actual construction site where the dust is too high and the light is insufficient,it is difficult to meet the processing requirements for the images taken directly.In this paper,the image denoising and enhancement algorithm is used to simply preprocess the captured images.For images with poor acquisition conditions and difficult quality requirements,an image interpolation and magnification algorithm based on triangular patches is proposed to supplement the image features.In order to solve the problem that the tunnel rock mass structure is diversified and difficult to deal with in a unified manner,this paper proposes a new crack classification method that divides the face cracks into strips and irregularities,using the improved Hough transform and the Livewire algorithm performs linear group extraction and magnetic tracking extraction respectively,finally clear and definite results are output.According to the result information,in-depth analysis is performed to obtain the quality criteria such as trace length,width,angle,center coordinates,and spacing of cracks by obtaining backbone pixels.Finally,a tunnel face image processing system is designed and implemented,and the research results are put into practical engineering applications.The tunnel face image processing system designed and implemented in this paper is put into use in a tunnel in Jinan City,Shandong Province.By controlling the dust concentration and lighting conditions,digital cameras and tungsten iodine lamps are used to capture images of tunnel face.Then they can be transmitted in real time and be processed.From the processing and extraction results of these images,the new image interpolation algorithm works well for low-resolution images to restore high-resolution images,providing more details for crack extraction on the face image.In the classification extraction stage,the results of crack extraction for different features are improved by the improved Hough transform and Livewire algorithm.The final analysis result is clear,and it fully meets the engineering requirements.