Abstract:
Riparian Vegetation monitoring using remote sensing images involves using satellite or aerial imagery to assess the health, extent, and changes in vegetation along water bodies such as rivers and streams. This monitoring is important to balance the aquatic ecosystems. The paper presents a novel methodology to detect Riparian Vegetation change dynamics using multi-temporal satellite imagery. The workflow performs the image pre-processing task that includes multi-temporal image registration, top-of-atmosphere reflectance computation, and NDVI geophysical parameters retrieved to quantify the amount and health of vegetation in a Riparian zone buffer. The changes in Riparian Vegetation are detected using the Kullback-Leibler (KL) divergence technique and a change detection map is generated. The proposed approach is evaluated using recent INS-2B Nano-MX and historical IRS-1C LISS-3 imagery near Sabarmati River, Gandhinagar, India. It is found from the visual assessment that there is an increase in healthy Riparian Vegetation around the Sabarmati River. The results are cross-validated with Landsat NDVI time series analysis in the Riparian Vegetation Region of Interest (ROI) for the last two decades and it is found that vegetation change estimated around the river stream closely follows the trend with the proposed method.