摘要
在脉冲涡流热成像检测中,有效抑制红外热图的噪声是最终提取特征量识别缺陷的关键环节之一。将提升小波阈值去噪的思想运用到二维经验模态分解(BEMD)中,提出了一种基于BEMD的提升小波阈值去噪方法。针对传统软、硬阈值法的局限性,引入包括带有可变因子的隶属函数的模糊阈值处理方法。将该方法运用于脉冲涡流热成像信号的实际消噪处理,实验结果表明,该方法与小波阈值去噪相比,去噪效果更明显,图像的细节特征更清晰。
Effectively suppressing the noise in infrared images is the key to extract features to identify the defects in Pulse Eddy Current(PEC) thermography detection. A new lifting wavelet threshold de-noising method is presented based on Bidimensional Empirical Mode Decomposition(BEMD). The fuzzy thresholding method which including a membership function with a variable factor was introduced to deal with the limitations of traditional soft and hard threshold method. The new de-noising method is applied to de-noise the pulse eddy current thermography signals. Compared to conventional wavelet threshold de-noising method, experimental results show that the de-noising performance of this method is better and the details of image feature are clearer.
出处
《红外技术》
CSCD
北大核心
2012年第6期346-350,共5页
Infrared Technology