Classification of some strategic crops in Egypt using multi remotely sensing sensors and time series analysis

Abstract

The agricultural fields in Egypt are commonly distributed with relatively small sizes parcels that usually reduce the reliability of Agricultural statistics in surveying cropland. The use of remote sensing help in an accurate crop inventory under complex landscape conditions based on the spectral characteristics differences of crops. The current study was carried out in Abu El Matamir district, Behira Governorate, located in western Nile delta Egypt. The main objective of the current study is using time series analysis of remote sensing data in crop discrimination. In this study, 160 locations of ground truth points collected during different growth stages of summer season crops. Two different sensors images used in this study represented by single date image of RapidEye and multi-date Landsat 8 OLI satellites. The acquired satellite images from both sensors atmospherically and geometrically corrected. Moreover different vegetation indices calculated such as Normalized Difference Vegetation Index (NDVI), Soil Adjusted Vegetation Index (SAVI) and Enhanced Vegetation Index (EVI) for cultivated crops in the study area during the whole growing season. Preliminary statistical analysis applied to the collected field data to show the distribution of the cultivated crops types. Moreover, unsupervised Iso-Data applied for multi-date Landsat 8 OLI images and calculated VI’s series for overall growth season. Results showed higher overall kappa accuracy with 0.82 and 0.79 respectively. NDVI showed the best representation of the crop phenological changes during the crop growth season and showed higher accuracy in strategic crops discrimination than the single date image with higher spatial resolution. Keywords: Vegetation index, Crops discrimination, Time series analysis, Iso-data classification