Classification of Hyperspectral Images by Using Extended Morphological Attribute Profiles and Independent Component Analysis
read more
Citations
Advances in Spectral-Spatial Classification of Hyperspectral Images
Spectral–Spatial Feature Extraction for Hyperspectral Image Classification: A Dimension Reduction and Deep Learning Approach
Spectral-spatial classification of hyperspectral imagery with 3D convolutional neural network
Spectral–Spatial Hyperspectral Image Classification With Edge-Preserving Filtering
Deep learning classifiers for hyperspectral imaging: A review
References
Independent Component Analysis
Visual pattern recognition by moment invariants
Classification of hyperspectral data from urban areas based on extended morphological profiles
Spectral and Spatial Classification of Hyperspectral Data Using SVMs and Morphological Profiles
Related Papers (5)
Classification of hyperspectral remote sensing images with support vector machines
Frequently Asked Questions (11)
Q2. What classes were found in the center area?
In the center area, the thematic classes found were Water, Tree, Meadow, Self-blocking Bricks, Soil, Asphalt, Bitumen, Tile, and Shadow.
Q3. What were the thematic classes identified in the city of Pavia?
Nine thematic land-cover classes were identified in the university campus: Trees, Asphalt, Bitumen, Gravel, Metal sheets, Shadows, Selfblocking Bricks, Meadows, and Bare soil.
Q4. How many samples were used for the training set?
3.In the analysis carried out, all the samples of the training set were used for the University data set, while for the Center data sets, only 50 samples (randomly chosen from the full training set for each class) were considered.
Q5. What was the method used to select the components of the EAP?
The model selection in the training phase of the classifier was based on a gradient descent method, which proved to be computationally less demanding than the exhaustive investigation of the parameters on a grid approach, giving comparable results [12].
Q6. What is the OA obtained by considering the EAPs?
When looking at the performances obtained by considering the spatial features extracted by the EAPs, one can see that the EAP with area attribute outperformed the other single EAPs with PCA, while when considering the ICA, the choice of the standard deviation performed the best among the single EAPs.
Q7. How many EAPs were computed by considering the attributes of the components extracted by PCA?
Four EAPs were computed by considering four different attributes on the components extracted by PCA and ICA: 1) a, area of the regions (λa = [100 500 1000 5000]); 2) d, length of the diagonal of the box bounding the region (λd = [10 25 50 100]); 3) i, first moment invariant of Hu, moment of inertia [11] (λi = [0.2 0.3 0.4 0.5]); and 4) s, standard deviation of the gray-level values of the pixels in the regions (λs = [20 30 40 50]).
Q8. What is the OA obtained by using the EAPs?
The best OA obtained by using the EAPs is higher, of about 2%, than those obtained by the original spectral features and the first components.
Q9. What is the significance of the spatial information in the data set?
For this data set also, it is evident the importance of including the spatial information, which led to an increase in terms of accuracy with respect to considering the original hyperspectral data or the components obtained from the dimensionality reduction technique.
Q10. What was the purpose of the experiment?
The experimental analysis was carried out on two hyperspectral images acquired over the city of Pavia (Italy) by the ROSIS-03 (Reflective Optics Systems Imaging Spectrometer) hyperspectral sensor.
Q11. What is the case for the FA?
The FA is performing well in average and has a robust behavior since, in all the experiments, the accuracies obtained, when compared to those of the single EAPs, are slightly lower than the best case (less than 2% of OA) and better than all the others.