| Most of the common dam diseases begin to appear on the surface of the structure,such as cracks,leakage,damage,and external deformation of the structure.Among them,crack is the most common form of disease.The crack monitoring method based on machine vision is simple and feasible.It can quickly extract crack feature information,eliminate subjective errors of manual measurement,improve the accuracy and efficiency of crack detection,and reduce costs.In this paper,a monitoring system based on digital image processing technology is used to collect disease images,and the methods of color image processing,gray image preprocessing and image segmentation in crack detection algorithm are studied.An image segmentation algorithm based on salp swarm algorithms is proposed,which is of great significance to dam crack image detection.The following work has been accomplished in this paper:(1)This paper establishes the concrete dam crack recognition system based on digital image processing technology.It mainly includes selecting the civil network cameras(DS-2DC6223IW-A and DS-2CD3T46DWD-i5)as the image acquisition equipment,and building concrete walls in the laboratory to simulate the hydraulic dam.(2)This paper studies crack image processing methods,including color image processing and gray image processing.The region growing method is effective in color image segmentation,but it has some subjectivity and limitations.The crack gray image segmentation methods include Canny,iteration method,Otsu method,particle swarm optimization algorithm and salp swarm algorithm.The SSA method proposed in this paper has the advantages of short time,small computation and good segmentation effect.(3)This paper introduces the feature information extraction method of crack image,and establishes a concrete dam disease monitoring GUI system based on MATLAB.(4)Application example analysis.This paper analyzes an engineering example according to the collected image of a reservoir cracks.It verifies the feasibility of crack detection algorithm and practicability of the system GUI system. |
PDF Full Text Request