CONTINUOUS ANOMALY DETECTION IN SATELLITE IMAGE TIME SERIES BASED ON Z-SCORES OF SEASON-TREND MODEL RESIDUALS

Natural disasters or human activities, such as forest fire, flood, and deforestation, may lead to anomaly or disturbance of land cover. Continuous detection of anomalies is important for studying spatial-temporal processes of land cover changes. Although many time-series-analysis methods have been developed for change detection, to the best of our knowledge, few methods focus on continuously detecting anomalies in satellite image time series. This study proposes a method for continuous anomaly detection in satellite image time series based on Z-scores of Season-Trend model Residuals (ZSTR). The ability of ZSTR for continuous anomaly detection was validated with an experiment for detecting spatial-temporal anomaly regions caused by severe flooding.