基于纹理特征的红外图像溢油检测算法研究  

Oil Spill Detection by Infrared Images Based on Texture Features

作　　者：阮沁馨 田金文 RUAN Qinxin;TIAN Jinwen(National Key Laboratory of Science and Technology on Multi-spectral Information Processing Technology,School of Automation,Huazhong University of Science and Technology,Wuhan 430074)

机构地区：[1]华中科技大学自动化学院多谱信息处理技术国家级重点实验室,武汉430074

出　　处：《计算机与数字工程》2018年第4期797-801,共5页Computer & Digital Engineering

基　　金：国家自然科学基金项目(编号:61273279)资助

摘　　要：针对船载红外相机拍摄海面红外溢油图像的特点,提出了一种基于纹理特征的溢油区域识别算法。首先利用面向对象的超像素分割方法,将红外溢油图像分割成若干个子区域,然后对每个子区域提取纹理特征,采用支持向量机分类器对海面和油膜纹理特征进行分类,建立红外溢油图像溢油区域识别的数学模型。在实验过程中,比较了采用不同纹理特征集的分类效果,并与传统分割算法进行比较,实验结果表明论文算法能较好地识别溢油区域,并保存溢油区域的边缘信息。According to the characteristics of infrared oil spill images which are captured by ship borne infrared camera,a recognition algorithm based on texture feature is proposed.First the object oriented superpixel segmentation is used,the infrared oil spill image is divided into several sub regions,next extract the texture features of each sub region,and the Support Vector Machine(SVM)is used to classify the texture features of sea surface and oil films,then the mathematical model for oil spill area identification by infrared oil spill image is established.In the experiment,the classification results of different texture feature sets are compared,and the result is compared to the traditional segmentation algorithm,the experimental results show that this algorithm can better identify the oil spill area and save the edge information of the oil spill area.

关 键 词：超像素分割 纹理特征 溢油 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]