Displacement Monitoring Of Bridges Based On Image Processing Method

Nowadays, the technology of image matching develops rapidly. More and more algorithms reach the state of art. However, a common problem which hasn't been solved well so far is the high computational cost. It means that it is hard to balance the matching precision and matching speed. Thus, a high precision matching algorithm with a faster speed and robustness is the goal of the researchers. In order to monitor the shaking of the bridge in real time, an advanced image matching algorithm is proposed in this paper.This paper firstly introduces the significance of this research. The research status is also given. Next, some well-known image matching algorithms and searching strategies are briefly introduced. The image matching methods can be grouped into two groups. One is using feature of images, the other is applying the info of gray values in images. This paper emphasizes several well-known algorithms which are using the info of the gray values, for instance, SSDA, ABS, NCC and so on. The advanced algorithm is based on NCC and optimizes the searching strategy. The optimized searching strategy combines hierarchical search with hill-climbing search, which divides the whole searching procedure into ‘Matching in rough' and ‘Matching in precise'. ‘Matching in rough' is to locate the object in a rough area, and the ‘Matching in precise' is to find the optimal location. For details, ‘Matching in rough' contains equivalence graph extraction step and hill-climbing search step, ‘Matching in precise' searches in a pixel-wise fashion in order to find the optimal matching result. In order to monitor the shaking of the bridge, the method in this paper separates into two parts. One is one frame image matching using both ‘Matching in rough' and ‘Matching in precise'. And the other is image matching based on former frames just using ‘Matching in precise'.A method in monitoring the shaking of the bridge based on the Monte Carlo particle filter is also introduced. A comparison between this and our method is shown to illustrate the performance. Last but not least, the whole implement of the image matching algorithm and the design of the monitoring bridge shaking system are give in the final of this paper.