Abstract

Identification and mapping of natural vegetation are major issues for biodiversity management and conservation. Remote sensing monitoring has two important sub fields viz, classification and change detection. Change detection from remotely sensed images is a process that utilizes the images acquired over the same geographical area at different times to identify the changes that may have occurred between the considered acquisition dates. Remotely sensed data with very high spatial resolution are currently used to study vegetation, but most satellite sensors are limited to four spectral bands, which is insufficient to identify some natural vegetation formations. To address this issue, this work intends to use multispectral and high spatial resolution data (Sentinel-2) to do the task of natural vegetation analysis using a popular CNN architecture. This proposed project concept is completely based on Deep Learning architecture using Remote Sensing technology for vegetation of lands in Hassan city of Karnataka India. The results are discussed within the scope of recent studies involving machine learning and Sentinel-2 data and key knowledge gaps identified. The changes of various vegetations are compared and their changes difference is displayed.