图/表检索结果

更多选择

期刊	DOI

起始年	结束年

卷	期

作者	作者单位

检索结果

期刊

Loading ...

出版年

Loading ...

Table 3 A detailed comparative analysis of the different segmentation methods
Figure 19 SAM-based UF extraction result diagrams, the extracted urban areas show the original colour that distinguishes them from non-urban areas, with a blue border distinguishing the two types of environments. Among them, Images a and b are from Beijing, China, c and d are from New Delhi, India, e and f are from Mansouria, Egypt, and g and h are from Porto Alegre, Brazil. In the legend, i and j from New York, USA, k and l from Kisangani, Sudan, m and n from London, UK, o and p from Phan Thiet, Vietnam.
Table 2 Summary of Kappa precision values for UF extraction based on SAM
Figure 18 Comparison of UF extraction details between the two deep learning algorithms. Graph a shows that U-Net misclassifies some of the urban shadows, and graph b shows that DeeplabV3 misclassifies some of the smaller inner-city public green spaces. The circle represents the misjudged area.
Figure 17 The results of UF extraction based on two deep learning algorithms, are presented for images i through p. Among them, Images i and j are from New York, USA, k and l from Kisangani, Sudan, m and n from London, UK, o and p from Phan Thiet, Vietnam. In the legend, 1 represents urban areas and 2 represents non-urban areas.
Figure 16 a to h images of UF extraction results based on two deep learning algorithms. Among them, Image a and b are from Beijing, China, c and d are from New Delhi, India, e and f are from Mansouria, Egypt, and g and h are from Porto Alegre, Brazil. In the legend, 1 represents urban areas and 2 represents non-urban areas.
Figure 15 Kappa precision numerical curve of UF extraction based on deep learning
Figure 14 Partial classification misclassification region visualization results (object-based)
Figure 13 Object-based Kappa precision numerical graph for UF extraction based on four traditional algorithms
Figure 12 Partial classification misclassification region visualization results (pixel-based)
Figure 11 The results of UF extraction based on four traditional algorithms, both pixel-based and object-based, are presented for images i through p. Among them, Image i and j from New York, USA, k and l from Kisangani, Sudan, m and n from London, UK, o and p from Phan Thiet, Vietnam.
Figure 10 a to h images of UF extraction results based on four traditional algorithms, both pixel-based and object-based. Among them, Image a and b are from Beijing, China, c and d are from New Delhi, India, e and f are from Mansouria, Egypt, and g and h are from Porto Alegre, Brazil.
Figure 9 Pixel-based UF extraction Kappa precision numerical graph based on four traditional algorithms
Figure 8 Correlation analysis diagram of UF extraction accuracy (Kappa) with different algorithms
Figure 7 Kappa accuracy values for all images under different classification methods
Figure 6 Schematic diagram of the classification principle of SVM
Figure 5 Segmentation results based on the Mean Shift algorithm (a. original image; b. segmentation result)
Figure 4 Illustration of an example of sample selection for UF extraction (a-j represent the layout of different cities and the selection of samples. a and b correspond to Beijing, c to New Delhi, d to Mansouria, e to Porto Alegre, f and g to Kisangani, h and i to London, and j to New York.)
Figure 3 Flowchart of the operation of this experiment
Table 1 Detailed description of the characteristics of different cities

跳至
页
第1页
共7649页
共152980条记录