摘要
为了识别信息量较少的红外图片,采用SURF算法提取特征并匹配。首先进行滤波、增强和分割等图像预处理工作,然后通过SURF算法提取图像特征并识别匹配,最后经过RANSAC算法筛选,获得正确率较高的匹配点。实验表明,在红外图像识别过程中,采用合适的预处理算法以及鲁棒性较强的SURF算法,可以取得令人满意的识别效果。
In order to recognize infrared images which have less information,SURF algorithm was used to extract features and match them.First,the images were preprocessed with filtering,enhancing and segmenting.Then,the image features were extracted by the SURF algorithm and the matched.Finally,the RANSAC algorithm was adopted to get the correct matching points.The experiments showed that in the process of infrared image recognition,a suitable preprocessing algorithm and a robust SURF algorithm would achieve a satisfactory recognition results.
作者
赵飞
朱琪
潘丽杰
徐本亮
ZHAO Fei;ZHU Qi;PAN Lijie;XU Benliang(Department of Advanced Manufacturing Technology,Zhejiang Yangtze Delta Region Institute of Tsinghua University,Jiaxing 314000,China)
出处
《电工技术》
2018年第16期118-120,123,共4页
Electric Engineering
基金
集成自主探测器的分析型工业热像仪开发和应用(2013YQ470767)
关键词
图像处理
红外图像
SURF
识别
image processing
infrared image
SURF
recognition
