摘要
随着科技水平的提高,深度学习算法的出现为高分辨率遥感图像的研究带来了新的突破,但国内对于将深度学习应用于遥感图像处理的研究尚未广泛开展。为填补此类空白,提出一种基于卷积神经网络(CNN)的对于高分辨率多光谱遥感图像进行自动分类的方法,对传统CNN框架进行一定的优化并加入Inception结构,进而横向比对其与支持向量机(SVM)分类算法的实际分类效果。以卫星拍摄的地面实物图片为例对该方法进行实验,结果表明,所提出的基于CNN的分类方法相比于传统方法在精度上有显著提升,纹理特征更加突出,分类效果更加出众。
With the improvement of science and technology, the appearance of deeping learning algorithm has brought a new breakthrough to the research of high-resolution remote sensing images, but the research of applying deeping learning to remote sensing image processing in China has not yet been widely carried out. In order to fill this gap, a method of automatic classification of high-resolution multispectral remote sensing images based on convolutional neural network(CNN) is proposed. The traditional CNN framework is optimized to a certain extent and the concept structure is added, and then the actual classification effect with support vector machine(SVM) classification algorithm is compared horizontally. Taking the ground real pictures taken by satellite as an example, the experiment results show that the proposed CNN-based classification method has significantly improved accuracy, more prominent texture features and better classification effect compared with the traditional method.
作者
李玉峰
林辉
LI Yufeng;LIN Hui(School of Electronic Information Engineering, Shenyang Aerospace University, Shenyang 110136, China)
出处
《微处理机》
2019年第1期43-48,共6页
Microprocessors
基金
高分专项省域产业化应用项目(70-Y40G09-9001-18/20)
辽宁省教育厅基本科研重点项目(L201701)
作者简介
李玉峰(1969—),男,吉林省德惠市人,博士后,教授,主研方向:图像处理。
