| 摘要: |
| 为提高遥感影像融合质量,提升资源一号(ZY-1 02D)高光谱遥感影像滨海湿地植被分类精度,提出将ZY-1 02D高光谱影像与空间分辨率为10 m的哨兵2号(Sentinel-2)影像进行Brovey融合,并通过搭建AlexNet卷积神经网络对ZY-1 02D高光谱影像和Brovey融合影像的滨海湿地植被进行分类,与支持向量机、随机森林和BP神经网络分类算法进行精度对比。研究结果表明:经Brovey融合后,AlexNet、支持向量机、随机森林和BP神经网络算法的植被分类总体精度分别提高15.60%、7.00%、14.80%和10.00%,Kappa系数提高了21.35%、9.93%、18.97%、12.85%;基于Brovey影像融合与AlexNet算法的植被分类精度最高,总体精度为92.40%,Kappa系数为89.42%。空谱融合配合AlexNet卷积神经网络有效解决了高光谱遥感影像在滨海湿地植被分类应用中精度较低的问题,为滨海湿地植被资源动态监测提供技术和方法支撑。 |
| 关键词: ZY-1 02D 滨海湿地 Brovey影像融合 植被分类 AlexNet算法 |
| DOI：10.11759/hykx20220614001 |
| 分类号:TP751 |
| 基金项目:国家自然科学基金项目(41506106);江苏省自然科学基金项目(BK20221397) |
|
| Vegetation classification combining spatial–spectral feature fusion based on remote sensing and AlexNet algorithm in a coastal wetland |
|
XU Chen1, LU Xia2, SANG Yu1, HE Shuang1, LIU Jing-xuan1
|
|
1.School of Marine Technology and Geomatics, Jiangsu Ocean University, Lianyungang 222005, China;2.School of Geography Science and Geomatics Engineering, Suzhou University of Science and Technology, Suzhou 215009, China
|
| Abstract: |
| The spatial structure information of ZY-1 02D hyperspectral remote sensing image loses a lot, which seriously affects its accuracy in coastal wetland vegetation classification. This paper proposed the application of Brovey fuse ZY-1 02D hyperspectral image with that of Sentinel-2 at a spatial resolution of 10 m and classified the coastal wetland vegetation of ZY-1 02D hyperspectral and Brovey fusion images by building an AlexNet convolution neural network. The accuracy of the classification algorithm was compared with that of the support vector machine, random forest, and back propagation neural network. The results showed that after the Brovey fusion, the overall accuracy of vegetation classification of AlexNet, support vector machine, random forest, and back propagation neural network was improved by 15.60%, 7.00%, 14.80%, and 10.00%, respectively, and the Kappa coefficient was improved by 21.35%, 9.93%, 18.97%, and 12.85%, respectively. The accuracy of vegetation classification based on Brovey fusion and AlexNet was the highest, with an overall accuracy of 92.40% and a Kappa coefficient of 89.42%. Space spectrum fusion and AlexNet convolution neural network effectively resolved the limitations of low accuracy of hyperspectral remote sensing image in the application of coastal wetland vegetation classification and provided technology and method support for the dynamic monitoring of coastal wetland vegetation resources. |
| Key words: ZY-1 02D coastal wetland Brovey image fusion vegetation classification AlexNet algorithm |
