| 摘要: |
| 深度卷积神经网络(Deep Convolutional Neural Network, DCNN)模型虽已应用于遥感影像分类的算法研究,但普遍集中于城市区域,在滨海湿地多光谱影像分类的工作相对匮乏。本文基于2018年10月份黄河口入海两侧的LANDSAT8 OLI影像,提取植被指数和缨帽变换分量共9维光谱特征,构建融合浅层特征的8层DCNN分类模型,开展互花米草遥感监测的方法研究,并从不同的浅层特征来具体分析互花米草的监测结果。结果表明：1)在分类方法上,DCNN模型的总体分类精度最高,达到90.33%,与支持向量机(Support Vector Machine, SVM)、随机森林(Random Forest, RF)分类器相比,精度分别提高4.78%、2.7%,互花米草的生产者精度分别提高了2.56%,0.47%,说明在滨海湿地遥感影像分类中,DCNN有着更好的应用潜力；2)融合浅层特征后,DCNN的总体分类精度和互花米草的识别精度分别提高了0.34%和3.25%,有效提高了对互花米草的监测能力。其中,融合归一化植被水分指数(NDII)浅层特征的DCNN分类方法中,互花米草的识别精度提高最多,为2.56%,比值植被指数(RVI)次之,为2.32%。 |
| 关键词: 深度卷积神经网络 浅层特征融合 湿地分类 互花米草 黄河口 |
| DOI：10.11759/hykx20190130001 |
| 分类号:P237 |
| 基金项目:国家自然科学基金项目(61601133, 41706209)；高分海岸带遥感监测与应用示范项目(41-Y30B12-9001-14/16) |
|
| Remote-Sensed monitoring of Spartina alterniflora using DCNN method with fusion of shallow features |
|
zhuyuling1, wangjianbu1, wangandong2, wangjinjin3,4, zhaoxiaolong1, renguangbo1, huyabin1,5, chenxiaoying1, mayi1
|
|
1.First Institute of Oceanography, Ministry of Natural Resources;2.Administration Bureau of the Yellow River Delta National Nature Reserve;3.Zhuhai Orbit Aerospace Science &4.Technology Co., Ltd;5.Dalian Maritime University
|
| Abstract: |
| Although Deep Convolutional Neural Network (DCNN) model has been applied to remote sensing image classification, it is generally concentrated in urban areas, and the work of multi-spectral image classification in coastal wetlands is relatively scarce. Based on the LANDSAT8 OLI imagery on both sides of the Yellow River estuary in October 2018, this paper extracts a total of 9-dimensional spectral features of vegetation index and the tasseled cap transformation, and constructs an 8-layer DCNN classification model with shallow features to identify the S.alterniflora. The results of S. alterniflora are analyzed from different shallow features. The results indicated that: 1) The DCNN model has the highest overall classification accuracy, reaching 90.33%. Compared with the Support Vector Machine (SVM) and the Random Forest (RF) classifier, the accuracy is increased by 4.78%. 2.7%, the producer’s accuracy of S. alterniflora increased by 2.56% and 0.47%, respectively, it shows DCNN has better application potential in coastal wetlands classification; 2) After the fusion of shallow features, the overall classification accuracy of DCNN and the recognition accuracy of S. alterniflora increased by 0.34% and 3.25%, respectively, which effectively improved the monitoring ability of S. alterniflora. Besides, in the DCNN classification method that combines the shallow features of normalized vegetation water index (NDII), the recognition accuracy of S. alterniflora increased the most, which was 2.56%, followed by the ratio of vegetation index (RVI), which was 2.32%. |
| Key words: Deep Convolutional Neural Network fusion of shallow feature wetland classification Spartina alterniflora Yellow River Estuary |