Kernel principal component and maximum autocorrelation factor analyses for change detection

Principal component analysis (PCA) has often been used to detect change over time in remotely sensed images. A commonly used technique consists of finding the projections along the eigenvectors for data consisting of pair-wise (perhaps generalized) differences between corresponding spectral bands covering the same geographical region acquired at two different time points. In this paper kernel versions of the principal component and maximum autocorrelation factor (MAF) transformations are used to carry out the analysis. An example is based on bi-temporal Landsat-5 TM imagery over irrigation fields in Nevada acquired on successive passes of the Landsat-5 satellite in August-September 1991. The six-band images (the thermal band is omitted) with 1,000 by 1,000 28.5 m pixels were first processed with the iteratively re-weighted MAD (IR-MAD) algorithm in order to discriminate change. Then the MAD image was post-processed with both ordinary and kernel versions of PCA and MAF analysis. Kernel MAF suppresses the noisy no-change background much more successfully than ordinary MAF. The ratio between variances of the ordinary MAF 1 and the kernel MAF 1 (both scaled to unit variance) calculated in a no-change region of the images is 140 corresponding to 21.5 dB. Kernel MAF analysis also outperforms both linear and kernel PCA here (not shown).