基于形态学和OTSU算法的红外图像降噪及分割  

Infrared Image Denoising and Segmentation Based on Morphology and Otsu Method

作　　者：徐超[1] 冯辅周[1] 闵庆旭[1] 孙吉伟[1] 朱俊臻[1] 

机构地区：[1]装甲兵工程学院机械工程系,北京100072

出　　处：《红外技术》2017年第6期512-516,共5页Infrared Technology

基　　金：无损检测技术教育部重点实验室(南昌航空大学)开放基金(ZD201529007)

摘　　要：在涡流热像技术中,图像处理是进行缺陷特征提取和识别的关键基础,而增强的图像有助于提高涡流热像技术的检测效果。利用多种常用的红外图像处理方法进行增强处理,旨在解决红外图像信噪比不高、缺陷对比度低等问题。首先分析了噪声来源,通过算术运算对涡流红外图像进行预处理,然后采用基于形态学权重的自适应算法进行形态学降噪,最后利用二维最大类间方差法(OTSU)对图像进行分割。定性和定量分析结果均验证了该方法的有效性和适用性,研究成果为红外图像的特征提取和缺陷识别奠定了方法基础。In eddy current thermography, image processing is the key basis of defect feature extraction and recognition, and the enhanced image is helpful to improve the detection effect. Therefore, a method com- bining three infrared image processing algorithms is introduced to enhance the infrared image in order to solve the problems of low SNR and contrast of infrared images. Based on analyzing the noise sources, ed- dy current infrared images are preprocessed by arithmetic calculation and the adaptive algorithm based on morphological weight is used to reduce the noise. Further, the images are segmented by 2D OTSU algo- rithm. The validity and applicability of the proposed method are both qualitatively and quantitatively veri- fied. This study aims to provide a basis for defect feature extraction and recognition.

关 键 词：涡流热像 红外图像 形态学降噪 图像分割 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]