| The paper focuses on the processing crack images of the based camera scan bridge crack monitor system(BCMSBVFS), the processing images deal with crack image to get features which used on real-time evaluation and prediction of bridge health and safety。The process image is the important part of the BCMSBVFS。Although there is lots of image processing technologies,the paper only focuses on image mosaic,Image denoising and edge detecting. One image mosaic is mainly use of the scanned image reconstruction of the entire image, the image denoising is mainly to reduce noise and the image edge detection get the important edge, through which the three parts together to obtain information on important characteristics of cracks in the bridge.In researching the current number of existing image processing methods, but these methods have some shortcomings themselves. This paper focused on studying the experience of recent years, Empirical Mode Decomposition (EMD) can decomposes a Signal into some IMFs. Using EMD in the image mosaic, image denoising, image edge detection, and has made some achievements. In the image mosaic,the paper use EMD method to decompose image,and then apply template match method ,this method can avoid noise effect. In the image denoising, using EMD to decompose the noised image, then get a series of IMFs, different compositions of IMF can do denoising. The paper also compares the method to mean method and median method, the result is that the method can replace mean filter and median method. In the edge detecting, EMD can be used, EMD also achieved certain effect.Java language and Unified Modeling Language(UML) are applied to design and development a set of EMD-based method of image processing system, which implements one-dimensional EMD signal processing demonstration, EMD-based image denoising, EMD-based image edge detection and EMD-based image mosaic. The system is used to the research of paper as experimental platform, but also can be used as a subsystem BCMSBVFS system. |
PDF Full Text Request