基于Gabor滤波器的多特征融合活动轮廓模型研究  

Research on multi-feature fusion activity contour modeling based on Gabor filters

作　　者：李更生 成波 刘相国[2] LI Gengsheng;CHENG Bo;LIU Xiangguo(School of Mathematics and Statistics,Ankang University,Ankang,Shanxi 725000,China;School of Mathematics and Big Data,Chaohu University,Hefei 238000,China)

机构地区：[1]安康学院数学与统计学院,安康725000 [2]巢湖学院数学与大数据学院,合肥238000

出　　处：《自动化与仪器仪表》2025年第2期10-13,共4页Automation & Instrumentation

基　　金：陕西省教育厅科学研究计划项目(23JK0280);安徽省高校自然科学重点研究项目(KJ2019A0683);安徽省高校优秀青年人才项目(gxyq2019082)。

摘　　要：针对红外图像的对比度较低以及红外图像的成像原理不同于可见光图像,使得红外图像呈现出较为模糊的视觉效果,这使得分割的难度加大。为了解决这些问题,提出一种多特征融合的红外图像分割算法。首先,设计不同方向和尺度的Gabor滤波器组,应用于红外图像以提取图像的纹理特征,获得多方向和多尺度的纹理特征,选择最优的特征。然后,在一维熵的基础上引入能够反映灰度分布空间特征的特征量来组成图像的二维熵,提取图像的空间信息。其次,将Gabor滤波器组提取的纹理特征、空间特征以及强度特征相结合,形成多特征融合的图像分割方法。最后,对基于Gabor滤波器的多特征融合的活动轮廓模型方法进行实验验证,并对其性能进行分析。结果表明,该方法能够准确有效地实现对红外图像的分割,具有良好的应用前景。For the low contrast of infrared images and the imaging principle of infrared images is different from that of visible light images,which makes infrared images show a more blurred visual effect,which makes the segmentation more difficult.In order to solve these problems,a multi-feature fusion infrared image segmentation algorithm is proposed.Firstly,a Gabor filter bank with dif-ferent directions and scales is designed and applied to the infrared image to extract the texture features of the image,obtain the multi-directional and multi-scale texture features,and select the optimal features.Then,based on the one-dimensional entropy,the fea-ture quantity that can reflect the spatial characteristics of the gray scale distribution is introduced to form the two-dimensional entropy of the image,and the spatial information of the image is extracted.Secondly,the texture features,spatial features and intensity fea-tures extracted by the Gabor filter bank are combined to form a multi-feature fusion image segmentation method.Finally,the active contour modeling method based on multi-feature fusion with Gabor filter is experimentally verified and its performance is analyzed.The results show that the method can accurately and effectively realize the segmentation of infrared images and has good application prospects.

关 键 词：特征融合 GABOR滤波器 活动轮廓 红外图像 

分 类 号：TN911.73[电子电信—通信与信息系统] TP202[电子电信—信息与通信工程]