The Regularized Iteratively Reweighted MAD Method for Change Detection in Multi- and Hyperspectral Data
read more
Citations
Advances in Hyperspectral Image and Signal Processing: A Comprehensive Overview of the State of the Art
Change Detection in Synthetic Aperture Radar Images based on Image Fusion and Fuzzy Clustering
A Deep Convolutional Coupling Network for Change Detection Based on Heterogeneous Optical and Radar Images
Learning Spectral-Spatial-Temporal Features via a Recurrent Convolutional Neural Network for Change Detection in Multispectral Imagery
Domain Adaptation for the Classification of Remote Sensing Data: An Overview of Recent Advances
References
The Elements of Statistical Learning: Data Mining, Inference, and Prediction
Numerical Recipes in C: The Art of Scientific Computing
An Introduction to Multivariate Statistical Analysis
Related Papers (5)
Automatic analysis of the difference image for unsupervised change detection
Frequently Asked Questions (12)
Q2. What is the order of the projection variates in the dimensionality reducing scheme?
The ordering of the projection variates in the dimensionality reducing regularization scheme is by some projection index (such as variance, autocorrelation, deviation from normality or other) rather than by wavelength.
Q3. How many groups do the authors use to avoid overlap?
in a regularization scheme which combines dimensionality reduction and curvature penalization, the authors choose 43 groupsavoid overlap between bands from HyMap’s four detectors the authors use the following three two-bands-only groups: spectral bands and ).
Q4. What is the way to reduce the dimensionality of the data?
Possible (near) singularities may also be remedied by means of principal component analysis (PCA), maximum autocorrelation factor (MAF), projection pursuit (PP) analysis or other dimensionality reducing projections applied to the variables at the two points in time separately before doing canonical correlation and MAD analysis.
Q5. What is the standard deviation of the no-change observations?
Assuming also independence of the orthogonal MAD variates the authors may expect that the sum of the squared MAD variates for pixel after standardization to unit variance approximately follows a distribution with degrees of freedom, i.e., approximately(12)The standardization should ideally be done by means of the standard deviation of the no-change observations.
Q6. What is the effect of the combined regularization scheme?
If the authors use this combined regularization scheme, the general transformation invariance may be lost depending on the choice of dimensionality reduction scheme.
Q7. What is the mean value of the original MAD and IR-MAD methods?
For the entire image, the authors see that both the original MAD and IR-MAD method give the mean value the authors expect, namely six which is the number of degrees of freedom.
Q8. What is the likely result of the regularized analysis?
Based on the visual inspection of Fig. 15 in which the authors see only three of the 126 original spectral bands, the authors see that several of the areas that seemto change are more likely to be characterized as change regions in the regularized analysis.
Q9. What is the projection indices for the three obtainable projections?
The obtained projection indices for all three (or two) obtainable projections are shown in Fig. 16 (the wavelength for a projection is chosen as the middle wavelength for the group if possible, if not the first wavelength is chosen).
Q10. How much work could be done on the regularization scheme with hyperspectral data?
Limited experience on the regularization scheme with hyperspectral data shows that more work could be done both on determining which and how many groups of spectral bands to choose in the dimensionality reducing projections, which projection index to choose, and on determining the regularization parameter and the matrix .
Q11. How many iterations of the image are there?
The authors see that after 30 iterations the weights assigned to the rightmost 512 384 no-change part of the image all remain close to one unlike the weights in the leftmost 512 128 potential change part.
Q12. What is the way to interpret the results of the iterated scheme?
In general, to interpret results from this and other types of change detection schemes it is recommended to perform simultaneous inspection and analysis of:• the change images, here the MAD variates and the change/no-change measures; • weight plots; • spectra for selected pixels; • results from clustering or classification of changes; • mean spectra for selected groups or clusters of pixels; • (per cluster) plots of correlations between original data andchange variates.