摘要
针对脉冲涡流红外无损检测中的红外图像噪声大对比度低、非均匀加热、目标难以检测的问题,提出了一种涉及图像背景估计、图像目标增强、降噪和阈值分割的红外图像综合处理算法。首先以鲁棒主成分分析(RPCA)算法为基础,将红外图像进行背景与目标的分离;针对传统RPCA对图像背景描述不足的缺点,引入了加权核范数来更好地描述图像背景;其次构建加权核范数最小化(WNNM)去噪模型,对目标图像进行去噪处理,增强图像对比度。最后对去噪后的目标图像进行阈值分割,得到目标信息。仿真实验结果表明,与传统RPCA和双边滤波算法比较,该方法对于红外图像的目标检测从主观视觉和数值指标上都具有更好的效果。
Aiming at the problem of high noise,low contrast,non-uniform heating and difficult target detection of infrared image in pulsed eddy current infrared nondestructive testing,this paper presents an infrared image synthesis processing algorithm involving image background estimation,image target enhancement,noise reduction and threshold segmentation.Firstly,based on the robust principal component analysis(RPCA)algorithm,the infrared image is separated from the background and the target.To make up the poor performance of traditional RPCA in describing backgrounds,the Weighted Nuclear Norm was introduced to better describe the image background.Secondly,a weighted nuclear norm minimization(WNNM)denoising model was constructed to denoise the target image to enhance image contrast.Finally,threshold segmentation is performed on the denoised target image to obtain target information.The simulation results show that compared with the traditional RPCA and bilateral filtering algorithms,the proposed method has better effects on subjective visual and numerical indicators for infrared image target detection.
作者
马烜
邹金慧
Ma Xuan;Zou Jinhui(Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China;Engineering Research Center for Mineral Pipeline Transportation of Yunnan Province,Kunming 650500,China)
出处
《电子测量与仪器学报》
CSCD
北大核心
2019年第7期137-144,共8页
Journal of Electronic Measurement and Instrumentation
基金
国家自然科学基金(61663017)资助项目
关键词
鲁棒主成分分析
加权核范数最小化
阈值分割
无损检测
robust principal component analysis
weighted nuclear norm minimization
threshold segmentation
nondestructive testing
