Change Detection and Image Time-Series Analysis 1

8Multitemporal Analysis of Sentinel-1/2 Images for Land Use Monitoring at Regional Scale

Andrea GARZELLI and Claudia ZOPPETTI

University of Siena, Italy

8.1. Introduction

Human activities such as urban and industrial development or agricultural practices can be classified as landscape disturbance and can be detected and tracked through remote sensing analysis at different temporal and spatial scales (Lambin et al. 2006).

Such an investigation allows us to detect new land cover types, and possibly states varying the destination of use, i.e. land use transitions. This goal can be pursued by following two distinct approaches: a classification-based change detection or a feature-/pixel-based change detection (Zhang et al. 2019). In the first case, the analysis first produces classification maps and then finds changes between labeled pixels or objects of the classification maps at different times: a change map will be produced by comparing available maps with the classification map of recently acquired remote sensing images. According to the second approach, processing is applied directly to remote sensing data to produce a reliable change map. This chapter presents an experimental study following a feature-based approach for building detection, time-tracking of a construction site and changed area estimation at a large, regional scale.

Building detection from satellite images was almost impossible a few decades ago due to the low spatio-temporal resolution of satellite images, ...