Land Use Classification

Abstract

In focusing on this research subject of crop classification within Mehsana (Gujarat), India, we shall address this important issue in depth. The primary goal of this research is to determine the sowing pattern that sees the highest accuracy in the crop classification and as a result provides valuable information for the monitoring of crops, production forecasting, and implementation of sustainable agricultural practices. One of the main things we use is the Google Earth Engine which is a source of processing and analyzing the datasets of very great volumes of satellite data. We employ the tools of Earth Engine to run complete tests and determine the level of accuracy of each proposed model to ensure that they capitalize on the differences between the different crop types.Here, the specific area of concern is crop discrimination, therefore, the purpose of this research is to classify and identified the crops grown in different areas of the Mehsana district that uses remote sensing data. This method characterizes the spectral properties which represent two-dimensional characteristic of land cover directly from remotely sensed data. Plants are reflecting different spectra depending on their species and so the models can be used for detecting various crops. The steps involved in the study are: The machine learning process through data acquisition, data pre-processing, feature extraction, machine learn- ing model training, model evaluation, and results analysis.