Research On Subway Tunnel Crack Detection Technology Based On Image Processing

In recent years, urban rail transit has become one of the important measures to solve traffic congestion with its high speed, high traffic density, high passenger flow and other characteristics. Its rapid development has brought great convenience to people. However, at the same time, a large number of urban rail transit facilities transition from the construction period to the curing period, bringing great challenges to the line safety testing.In various security detections, tunnel crack detection has been a technical difficulty. At present, tunnel crack detection in our country still bases on manual judgment and recording. It is subjective and takes a lot of time and manpower, cannot meet the requirements of the tunnel structure safety assessment. Crack detection method based on image processing can compensate for the lack of traditional detection methods, acquiring crack information efficiently and accurately as well as saving labor costs. It is an ideal technology to automatically detect cracks.The main work of this paper includes the following aspects:(1)Research on the basic methods of crack image processing, introduce commonly used algorithms for image preprocessing (gamma correction, smoothing, sharpening), image segmentation and feature extraction, also analyze the applicability and shortcomings of each algorithm.(2)Analyze the characteristics of the tunnel crack image, improve2-D Otsu fast-iterative algorithm according to these characteristics. In the improved algorithm, local threshold is used instead of global threshold, a correction factor△and a threshold ω are proposed. The improved algorithm can solve the impact brought by uneven background, and can separate the tiny cracks in the tunnel more accurately. The width of the smallest crack which can be detected is0.23mm.(3)Analysis crack detection algorithms for each step of image processing, and simulate with Matlab, select the most appropriate algorithm after simulation, and then propose a tunnel crack detection method based on image processing. This method is proved to have higher accuracy rate and lower undetected rate after applying in50tunnel images. It can meet the requirement of application.(4)Design and implement the subway tunnel crack monitoring system with JAVA. Data processing module extracts tunnel crack information basis on the tunnel crack detection method proposed in this paper, then import it into the database; Data management module implements the CRUD of crack data, classifies cracks according to the threshold, and predicts disease trends by contrasting the data to original data, implementing timely warning.Finally, this paper has tested the subway tunnel crack monitoring system, which not only can efficiently and accurately obtain tunnel crack data, but also realizes the intelligent management of these data. It meets the rail transit safety testing requirements.