Implementation of remote sensing for vegetation studying using vegetation indices and automatic feature space plot

Abstract

Remote sensing techniques play an important role for monitoring vegetation growth and health, as well as others Landcover and Landuse. Image segmentation techniques are the most important tools, usually used to differentiate between the Earth’s surface features. One of the most common techniques to isolating the vegetated area from other land use regions is by utilizing the vegetation indices.

In this research, different vegetation indices will be utilized for detecting and monitoring vegetation greenness, healthiness, and wetness.

A new adaptive technique for image segmentation has been introduced in this research is automatic feature space plot, this based on partitioning the feature space plot between the visible Red and Near-Infrared remotely sensed bands. The multi temporal Enhanced-Thematic-Mapper plus (ETM+) available scenes have been used to cover the studied areas, in two successive years (2001 and 2002). This feature space plot segmentation method divided the reflectance diagram in two regions; these were vegetation and no-vegetation.

A variety of indices formulas have also been used to globalize the vegetation patches, three of these vegetation indices have been adopted (i.e. RVI, NDVI and IPVI). The “NDVI” has higher recognized vegetated areas than other adopted indices of the amount of vegetation (ripe vegetation). Image binarization method being followed the implementation of the indices to isolating the vegetation areas from the image background. The isolated vegetated areas and their percentages are presented in tables to show the agriculture regions in two multi temporal scenes. The changes at these agriculture areas have also been computed and presented visually on the form of images, and numerically by listing them in tables (in km²). The counted areas resulted from the automatic feature space plot method and the isolated vegetated areas resulted from the implementation of the vegetation indices are also presented.

Keyword Image segmentation, feature space plot segmentation, vegetation indices, image binarization, change detection technique.