![]() extracted spectral features from one-dimensional CNN extracted local spatial features of each pixel from two-dimensional CNN, and further developed 3-D CNN to learn the spatial and spectral features of his. Recently, the 3D-CNN algorithm could take the spectrum as the third dimension to fully utilize the spatial-spectral information of hyperspectral data. How to make full use of spatial-spectral information is particularly important. With the development of remote sensing technology, the spatial and spectral resolution of hyperspectral data is getting higher and higher. Therefore, there is an urgent need for a method that could fully mine the spectral-spatial information of hyperspectral data, especially including the shadow areas. Additionally, most of the studies classified the urban objects in shadow areas separately. However, these studies shown that traditional classification methods could not make full use of the rich spatial-spectral information of hyperspectral data. improved the simple linear iterative cluster (SLIC) method, which showed better classification performance on three HIS compared with SVM. To better explore the spectral characteristics of HSI, Zhang et al. Compared with random forests and support vector machines, the features extracted by OTVCA show considerable improvement in classification accuracy. applied orthogonal total variation component analysis (OTVCA) to urban hyperspectral images with high spatial resolution. employed pixel-based support vector machines and object-based classifiers to extract urban objects in cloud shadows. separately classified the shaded and the unshaded areas and acquired the formal results by decision fusion with an overall accuracy of 95.92%. used hyperspectral data to classify all shaded pixels in urban areas with different land cover types using a maximum likelihood classifier (MLC) and SVM classifier. As for the shadowing problem, Qiao et al. ![]() Tamilarasi and Prabu have extracted roads and buildings in urban areas using SVM, and the achieved accuracy is 78.34% and 92.47%, respectively. utilized hyperspectral data and LiDAR data for the extraction of urban tree species using RF and achieved an overall accuracy of 87.00%. used hyperspectral data for urban feature extraction using SVM, and the overall accuracy is 90.02%. The appearance of hyperspectral images at different times with high resolution makes it possible to overcome those problems with more details and improve the accuracy. Compared with the summer hyperspectral index, the overall accuracy of the multitemporal hyperspectral index is improved by 0.9~3.1%. compared the effect of hyperspectral characteristics in different seasons on the mapping of land cover. ![]() drew an urban land cover map, and the overall classification accuracy reached 97.24%. Through high-resolution hyperspectral data to provide detailed structural information and spectral information, Chen et al. Urban feature extraction has always been a difficult and hot issue due to the problems of different things with the same spectrum patterns, the same things with different spectrums, shadows, and spatial heterogeneity. In the future, 3D-1D-CNN could also be used for the extraction of urban green spaces. The results indicated that 3D-1D-CNN could mine spatial-spectral information from hyperspectral data effectively, especially that of grass and highway in cloud shadow areas with missing spectral information. The overall accuracy of the proposed 3D-1D-CNN is 96.32%, which is 23.96%, 11.02%, 5.22%, and 0.42%, much higher than that of SVM, RF, 1D-CNN, or 3D-CNN, respectively. Finally, a confusion matrix and Kappa coefficient were calculated for accuracy assessment. Thirdly, Support Vector Machine (SVM), Random Forest (RF),1D-CNN, 3D-CNN, and 3D-2D-CNN classifiers were also carried out for comparison. Secondly, the parameters were fused and segmented into many S × S × B patches which would be input into a 3D-CNN classifier for feature extraction in complex urban areas. Firstly, spectral composition parameters, vegetation index, and texture characteristics were extracted from hyperspectral data. Therefore, a 3D-1D-CNN model was proposed for feature extraction in complex urban with hyperspectral images affected by cloud shadows. However, its capability of data mining in complex urban areas, especially in cloud shadow areas has not been validated. Recently, a three-dimensional convolutional neural network (3D-CNN) provides a new effective way of hyperspectral classification. However, how to mine and use this information effectively is still a great challenge. Airborne hyperspectral data has high spectral-spatial information.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |