Change Detection and Image Time Series Analysis, Volume 1: Unsupervised Methods

Change Detection and Image Time Series Analysis 1 presents a wide range  of unsupervised methods for temporal evolution analysis through the use  of image time series associated with optical and/or synthetic aperture  radar acquisition modalities.

Chapter 1 introduces two  unsupervised approaches to multiple-change detection in bi-temporal  multivariate images, with Chapters 2 and 3 addressing change detection  in image time series in the context of the statistical analysis of  covariance matrices. Chapter 4 focuses on wavelets and  convolutional-neural filters for feature extraction and entropy-based  anomaly detection, and Chapter 5 deals with a number of metrics such as  cross correlation ratios and the Hausdorff distance for variational  analysis of the state of snow. Chapter 6 presents a fractional dynamic  stochastic field model for spatio temporal forecasting and for  monitoring fast-moving meteorological events such as cyclones. Chapter 7  proposes an analysis based on characteristic points for texture  modeling, in the context of graph theory, and Chapter 8 focuses on  detecting new land cover types by classification-based change detection  or feature/pixel based change detection. Chapter 9 focuses on the  modeling of classes in the difference image and derives a multiclass  model for this difference image in the context of change vector  analysis.