IN-202541032668-A - ANALYZING LAND COVER CHANGES WITH LANDSAT SATELLITE DATA: AN APPLICATION TO ENSEMBLE LEARNING

IN202541032668AIN 202541032668 AIN202541032668 AIN 202541032668AIN-202541032668-A

Abstract

The present invention introduces a novel system for land cover change detection, integrating Landsat satellite data, ensemble learning techniques, and automated change analysis to enhance accuracy and scalability in environmental monitoring. The system leverages machine learning and AI-driven classification models, including Random Forest (RF), Gradient Boosting (GBM), Support Vector Machines (SVM), and Convolutional Neural Networks (CNNs), to improve land cover classification precision while minimizing misclassification errors. A key innovation of this invention is the automated multi-temporal change detection framework, which compares classified satellite images over different time periods using pixel-based, object-based, and deep learning models. The system applies radiometric correction, cloud masking, feature selection, and dimensionality reduction (e.g., PCA) to optimize input data and ensure high-quality classification. By integrating bagging, boosting, and stacking ensemble learning methods, the invention achieves greater accuracy and robustness compared to conventional single-classifier approaches.Furthermore, the invention incorporates a self-learning AI mechanism that continuously updates its classification models based on newly acquired satellite data, ensuring adaptability and long-term accuracy improvements. The results are presented in an interactive GIS-based visualization platform, enabling policymakers, urban planners, and environmental researchers to analyze land cover transitions, deforestation trends, urban expansion, and climate-induced changes effectively.This fully automated, scalable, and high-precision system revolutionizes land cover change detection by integrating remote sensing, machine learning, and AI-driven analytics, making it a vital tool for sustainable environmental monitoring, urban planning, and climate resilience research.

Inventors

DR MOHEBBANAAZ ASSOCIATE PROFESSOR DEPT OF ECE
MRS C AHALYA ASSOCIATE PROFESSOR DEPT OF ECE
MRS M JYOTHIRMAI ASSISTANT PROFESSOR DEPT OF ECE
MR K V SIVA REDDY ASSISTANT PROFESSOR DEPT OF ECE
MR P KISHOR KUMAR ASSISTANT PROFESSOR DEPT OF ECE
MR A RAJENDRA BABU ASSISTANT PROFESSOR DEPT OF ECE
MS S ISHRATH MOIN ASSISTANT PROFESSOR DEPT OF ECE
DR M JAYALAKSHMI PROFESSOR DEPT OF ECE

Dates

Publication Date: 20250425
Application Date: 20250402
Priority Date: 20250402

Description

Complete SpecificationDescription:The invention shown in Fig. 1 and Fig. 2 operates on a systematic workflow that integrates Landsat satellite data processing, machine learning-based classification, and ensemble learning techniques to accurately detect land cover changes. The core operational principle consists of the following key stages:1. Data Acquisition and Preprocessing