Detection of land cover classes in agro-ecosystems of northern Egypt by remote sensing

Abstract

Abstract The objective of this study is to detect the current land cover classes of a representative area in the coastal agricultural land of Egypt, where changes have been observed recently due to the urban sprawl. This has been achieved after the production of different vegetation indices, and the comparison between them and their sensitivity to the environmental conditions of the area under study. This was followed by the production of profiles of different Vegetation Index (VI) images, and supervised Maximum Likelihood Classification (MLC), for the assessment of land cover classes and their cover percentage. It was found that different greenness levels were best estimated by the Environmental Vegetation Index (EVI) and the Normalized Difference Vegetation Index (NDVI). The MLC resulted in classifying the study area into 10 classes with an accuracy above 90 per cent. Agriculture lands cover about 46 per cent only of the total area, which was estimated to cover at least 80 per cent 15 years ago. Land transformation is proceeding by a rate of about 1-47 per cent every year. This means that if such a rate continues in the same pattern, all agricultural land in the area will be lost to urbanization and other activities within less than 70 years.