Change detection in bi-temporal data by canonical information analysis

Canonical correlation analysis (CCA) is an established multivariate statistical method for finding similarities between linear combinations of (normally two) sets of multivariate observations. In this contribution we replace (linear) correlation as the measure of association between the linear combinations with the information theoretical measure mutual information (MI). We term this type of analysis canonical information analysis (CIA). MI allows for the actual joint distribution of the variables involved and not just second order statistics. Where CCA is ideal for Gaussian data, CIA facilitates analysis of variables with different genesis and therefore different statistical distributions. As a proof of concept we give a toy example. We also give an example with DLR 3K camera data from two time points covering a motor way.

[1]  H. Hotelling.  Relations Between Two Sets of Variates , 1936 .

[2]  Goldberg,et al.  Genetic algorithms , 1993, Robust Control Systems with Genetic Algorithms.

[3]  Masashi Sugiyama,et al.  Canonical dependency analysis based on squared-loss mutual information , 2011, Neural Networks.

[4]  F. Kurza,et al.  CALIBRATION OF A WIDE-ANGLE DIGITAL CAMERA SYSTEM FOR NEAR REAL TIME SCENARIOS , 2007 .

[5]  P. Reinartz,et al.  Calibration of a Wide-Angel Digital Camera System for Near Real Time Scenarios , 2007 .

[6]  D. E. Goldberg,et al.  Genetic Algorithms in Search , 1989 .

[7]  Paul A. Viola,et al.  Alignment by Maximization of Mutual Information , 1997, International Journal of Computer Vision.

[8]  Morton J. Canty,et al.  Image Analysis, Classification and Change Detection in Remote Sensing: With Algorithms for ENVI/IDL and Python, Third Edition , 2014 .

[9]  Allan Aasbjerg Nielsen,et al.  The Regularized Iteratively Reweighted MAD Method for Change Detection in Multi- and Hyperspectral Data , 2007, IEEE Transactions on Image Processing.

[10]  Yoav Y. Schechner,et al.  Fast kernel entropy estimation and optimization , 2005, Signal Process..

[11]  R. Fletcher,et al.  A New Approach to Variable Metric Algorithms , 1970, Comput. J..

[12]  Allan Aasbjerg Nielsen,et al.  Kernel principal component and maximum autocorrelation factor analyses for change detection , 2009, Remote Sensing.

[13]  L. Goddard.  Information Theory , 1962, Nature.

[14]  Xiangrong Yin,et al.  Canonical correlation analysis based on information theory , 2004 .

[15]  David J. C. MacKay,et al.  Information Theory, Inference, and Learning Algorithms , 2004, IEEE Transactions on Information Theory.

[16]  Robert B. Ash,et al.  Information Theory , 2020, The SAGE International Encyclopedia of Mass Media and Society.

[17]  Jacob S. Vestergaard,et al.  Canonical Information Analysis , 2015 .

[18]  John A. Nelder,et al.  A Simplex Method for Function Minimization , 1965, Comput. J..

[19]  Radford M. Neal.  Pattern Recognition and Machine Learning , 2007, Technometrics.

[20]  CameronIain.  Image analysis, classification and change detection in remote sensing , 2013 .

[21]  Allan Aasbjerg Nielsen,et al.  Kernel Maximum Autocorrelation Factor and Minimum Noise Fraction Transformations , 2011, IEEE Transactions on Image Processing.

[22]  T. Stephenson.  Image analysis , 1992, Nature.

[23]  Erkki Oja,et al.  Independent Component Analysis , 2001 .

[24]  Knut Conradsen,et al.  Multivariate Alteration Detection (MAD) and MAF Postprocessing in Multispectral, Bitemporal Image Data: New Approaches to Change Detection Studies , 1998 .

[25]  Leszek Wojnar,et al.  Image Analysis , 1998 .

[26]  R. A. Leibler,et al.  On Information and Sufficiency , 1951 .

[27]  Sang Joon Kim,et al.  A Mathematical Theory of Communication , 2006 .

[28]  Allan Aasbjerg Nielsen,et al.  Canonical analysis basedonmutual information , 2015, 2015 IEEE International Geoscience and Remote Sensing Symposium (IGARSS).

[29]  Peter Reinartz,et al.  Automatic traffic monitoring with an airborne wide-angle digital camera system for estimation of travel times , 2007 .