The Improved Algorithm Of Concrete Pavement Crack Detection Based On Percolation Model

With the rapid development of China's economy, there are more and more concrete highways and tunnels. The maintenance of the concrete surface is becoming improtant. In particular, crack on the concrete surface is one of the earliest indications of degradation. It will seriously damage the structure of concrete. The concrete surface crack detection method based on the digital image processing is much more effective than the manual inspection, such as simple detection equipment, low cost, convenient operation. It becomes popular crack detection technology.This thesis mainly research, the crack detection method based on the percolation model. First, the percolation algorithm has the features of high accuracy and low efficiency. To ensure a high efficiency, this thesis proposed a pre-extraction method to reduce the number of percolated pixels. Second, to ensure a high accuracy, this thesis proposed a second percolation processing algorithm to percolate the result of the first percolation processing, and reconnected the fractured cracks caused by missing detection. Finally, design and implement a concrete surface crack detection system. The main research contents include the following aspects:1. Researched the crack detection algorithm based on percolation model. To overcome the low detection efficiency of the percolation model. This thesis proposed a pre-extraction method to improve the efficiency of percolation. Divided image into several small grids with overlapping region. Then, extracted and merged the darker pixels from each grid. The experimental results show that the algorithm can detect the location and the basic shape of the crack.2. Proposed second percolation process to improve the accuracy of crack detection results. This method used second percolation process for the neighbourhood pixels which were marked as a crack by the first time percolation.3. Proposed connection algorithm to connect the fractured cracks. This method used hole filling, smoothing operation to process the image. Extracted the crack skeleton with the zhang fast parallel algorithm and the 8 direction chain code. According to the crack trend, analysed the fractured location.4. Designed and implemented a concrete pavement crack detection system. This system can process crack image by the method proposed in this thesis. Output binary image contains crack information to further research.