Hyperspectral remote sensing is a multi-dimensional information acquisition technology that combines imaging technology and spectral technology. It can obtain continuous and narrow band image data with high spectral resolution. Therefore, hyperspectral remote sensing has great potential in the identification of ground features and the classification of vegetation types. In this paper, GF-5 data was used as training data to classify forest types in Northeast China. Firstly, the water absorption bands and some noise bands were removed from the GF-5 hyperspectral image. Furthermore, the bands were grouped according to their correlation, and principal component analysis (PCA) was performed on each group of bands. According to the band index, the bands with better quality were extracted from each group and combined with the bands obtained by PCA to reduce the dimension of hyperspectral data. Then the Convolutional Neural Network (CNN) was used to extract the features of the processed image, and the extracted features were input into the support vector machine (SVM) classifier to obtain the forest vegetation type. By combining CNN and SVM, a hyperspectral forest classification model based on CNN-SVM fusion is constructed. The experimental results show that the method proposed in this paper performs best in forest type classification accuracy. The overall classification accuracy can reach 88.67%, and the Kappa coefficient can reach 0.84.
Land cover classification using UAV multi-spectral images is of great significance in precision agriculture, urban planning, land use and other fields. However, traditional remote sensing image classification methods cannot meet the classification accuracy requirements of UAV multi-spectral images. This paper aims to propose an object-based machine learning classification method to improve the land over classification accuracy of UAV multi-spectral images. The experimental area is a standard test field located in the Jilin Province of China. The experimental data was captured by a UAV equipped with a multi-spectral camera which includes four bands from 550 nm to 790 nm. First, the original images were preprocessed and the spectral curves of land cover were analyzed, thus four kinds of land cover with large differences were selected as categories. Then pixel-based, boosting-based and object-based machine learning methods were used for classification. The object-based classification method could make full use of the spatial and spectral information, and eliminate the noise problem caused by the high resolution of the UAV image to a certain extent. Finally, accuracy analysis using the verification image showed that the RF-O method achieved the highest classification accuracy of 92.2419%, and the kappa coefficient was 0.8904. All results indicate that the object-based machine learning classification method proposed in this paper is more suitable for the research of land cover classification, comparing with the traditional remote sensing image classification methods, and performs well on the land cover classification of UAV multi-spectral images.
The ground measurement of forest height is time-consuming and labor-consuming. In recent years, with the development of satellite remote sensing technology, it is possible to obtain forest height using radar remote sensing data. This paper uses the simulated full polarization radar data as the research object. The PCT inversion method (PCT), RVoG inversion method (RVoG), coherent amplitude inversion method (COH) and DEM difference inversion method (DEM) was used to obtain forest height. Through these four inversion methods, the influence of four parameters (radar incident angle, vegetation density, tree species, actual forest height) on the inversion result was studied. The experimental results show that if the radar incident angle is closer to 45 degrees, the vegetation density is greater than or equal to 300 trees, and the actual forest height is higher than 10 meters, the forest height inversion results have better accuracy. The research’s conclusions can provide a theoretical basis and method for error analysis of forest height inversion from real radar remote sensing data.
With the development of space satellite technology, a large number of high resolution remote sensing images have emerged. Deep learning has become an effective way to handle big data. Crop classification can estimate crop planting area and structure, and classification result is an important input parameter for crop yield model. Crop classification based on deep learning can further improve the estimation accuracy of production. In this paper, multi-temporal Sentinel-2 data and GF2 data are used as data sources. Sentinel-2 data is used as training data, and GF-2 data is used as validation data. Jilin Province in Northeast of China is selected as the experimental area. The experimental area is classified as rice, towns, corn and soy. Firstly, the multi-temporal Sentinel-2 data and GF-2 data is preprocessed. Then, the Sentinel-2 data is used to classify crops based on convolutional neural network (CNN) and visual geometry group (VGG). The red edge band, multiple indexes including normalization difference vegetation index (NDVI), normalized difference water index (NDWI) and difference vegetation index (DVI) are added respectively to compare with the classification results of original multitemporal Sentinel-2 data. The final classification results using CNN and VGG are compared with the other two machine learning algorithms including support vector machine (SVM) and random forest (RF). The experimental results show that the VGG performs best in crop classification accuracy. The overall classification accuracy of the crop can reach 94.8878%, and the Kappa coefficient can reach 0.9253, which is superior to the two traditional machine learning algorithms.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.