Abstract: This talk will be a brief compilation of different research publications/works which presents a diverse array of innovative methodologies and applications in the fields of remote sensing data analysis for environmental monitoring and agriculture. Notable contributions include a novel Mutual Information-based Segmented Stacked Autoencoder approach for hyperspectral data classification, a comprehensive evaluation of Feature Selection and Feature Extraction Techniques on Multi-Temporal Landsat-8 Images for enhanced crop classification accuracy, and the introduction of a Partial Informational Correlation Based Band Selection strategy for improved hyperspectral image classification. Additionally, studies on predicting canopy-averaged chlorophyll content using Convolutional Auto-Encoder, an evaluation framework for Landsat 8-based Actual Evapotranspiration Estimates in data-sparse catchments, and the transformation of multispectral data to quasi-hyperspectral data using Convolutional Neural Network Regression offer valuable insights into advancing utilization of machine/deep learning algorithms for agricultural/remote sensing applications. The compilation also delves into topics such as the spatio-temporal dependency of vegetation dynamics on climatic variables, flood concepts, and challenges, alongside an unsupervised annual change detection methodology from optical-SAR fused satellite image time-series using 3D-CAE, collectively contributing to the forefront of research in these domains. Current problem statement also deals with monitoring of regenerative agricultural practices using remote sensing data.

Meeting ID: 944 2649 3601
Passcode: 070134