Analysing Phenological Characteristics Extracted from Landsat NDVI Time Series to Identify Suitable Image Acquisition Dates for Cannabis Mapping in Afghanistan

作者：Mattiuzzi Matteo^*; Bussink Coen; Bauer Thomas

来源：Photogrammetrie Fernerkundung Geoinformation, 2014, 2014(5): 383-392.

DOI：10.1127/1432-8364/2014/0231

• illicit crop monitoring
• cannabis
• NDVI
• phenology
• seasonality
• senescence

Since 2009, the United Nations Office on Drugs and Crime (UNODC) has carried out various annual surveys of cannabis cultivation in Afghanistan. The mapping of cannabis fields is based on very high resolution satellite imagery. The image acquisition dates have a strong influence on whether cannabis can be visually differentiated from other crops. Expert knowledge shows that the ideal acquisition dates are at the end of the growing period (senescence) of cannabis. For future surveys a method was developed to predict the optimum observation dates independent of temporal differences of the phenology. The study includes an analysis of Landsat-based NDVI time series and the calculation of information about the senescence stages of cannabis and other vegetation. The approach was applied to four test sites in Afghanistan.

• 出版日期2014