基于SVM和相干平均去噪的大坝位移检测方法

在测量大坝、桥梁等建筑物的形变过程中，常需要采用微小位移测试技术．传统的测试方法难以实现在成本和维护费用较低、定标过程较为简单的前提下保证较高的测量精度．针对此问题，提出了一种新的基于支持向量机和相干平均去噪的大坝微小位移检测方法．首先通过自行研制的图像采集卡获取测量所需图像，然后用相干平均对图像进行预处理，最后用支持向量机进行进一步去噪，试验结果证明了该方法可以在低成本、低维护费用、定标过程较为简单的情况下使测量精度满足实际要求．

Displacement Detection of Dams Based on SVM and Coherence Average Denoising

When measuring the deformation of dams, bridges and other buildings, the micro-displacement detection technology is often used. However, with the traditional measurement method, it is very difficult to ensure higher measurement accuracy with lower cost and maintenance cost and simpler calibration. A novel approach to micro-displacement detection of dams based on Support Vector Machine and coherence average denoising is proposed. Firstly, the experimental image is captured by an image sample-card developed by the authors;secondly,the image is preprocessed by coherence average;and at last the Support Vector Machine is applied to denoise the preprocessed image. Experimental results show that the proposed method can ensure measurement accuracy with lower cost and maintenance cost and simpler calibration process.