基于红热图像测温技术的图像二维无网格消噪方法  

2D Meshless De-noising Method of Images Based on the Thermal Infrared Image Temperature Measurement Technology

作　　者：张玉存[1] 侯新玉 林洪彬[1] 

机构地区：[1]燕山大学电气工程学院

出　　处：《计量学报》2015年第4期379-383,共5页Acta Metrologica Sinica

基　　金：河北省交通运输厅科技计划项目(Y-131114)

摘　　要：红外热图像测温过程中存在的大量噪声严重制约了红外热图像的精度和准确性，为了准确提取红外热图像中被测物体的相关参数，提出了一种红外热图像二维无网格消噪方法。将双边滤波、异性消噪、无网格消噪进行结合，建立红外热图像二维无网格消噪模型，利用鲁棒回归算法求解模型中相关参数，实现了红外热图像的无网格消噪，提高了红外热图像测温的精度。最后，通过仿真实验验证了提出方法的可行性，为军事、科研等领域的红外热图像的测温提供了理论基础。The temperature measurement process of infrared thermal image has a large amount of noise, which has seriously hampered the precision and accuracy of infrared thermal image. In order to accurately detect the correlative parameters of the measured object in the infrared thermal image, a 2D meshless denosing method based on infrared thermal images is proposed. Bilateral filtering, anisotropic denoising, meshless denoising are combined to create a 2D meshless denoising model. The model parameters are estimated by using robust regression algorithm, and realize infrared thermal images with meshless denoising. And it enhances the temperature measurement precision of infrared thermal image. Finally, the feasibility of the proposed method is verified by simulation experiments and the method lays the theoretical foundation for military, scientific research and other areas of temperature measurement of infrared thermal image.

关 键 词：计量学 红外热图像 鲁棒回归算法 双边滤波 异性消噪 无网格消噪 

分 类 号：TB942[一般工业技术—计量学]