Image resampling: Nearest neighbor, bilinear and cubic convolution
Image mosaicking
Image normalization and indices
Image re-projection
Interpretation of Satellite Images and Image Statistics (8 hours)
Image interpretation: Techniques for visual and contextual analysis, and feature identification
Methods of image enhancement (Image clarity, stretching)
Classification: Methods for assigning pixel values to categories and object-based analysis
Descriptive statistics: Calculation and interpretation of mean, median, mode, and standard deviation of pixel values
Histogram analysis: Analysis of pixel intensity distribution through histograms
Digital Image Classification and Accuracy Assessment (10 hours)
Classification methods (Supervised and unsupervised approaches)
Supervised classification techniques: Maximum likelihood, Support Vector Machines (SVM), XGBoost and Random Forest
Unsupervised classification techniques: Clustering methods (k-means and ISODATA)
Post-classification refinement process
Sampling techniques: Ground truth data collection methods (Random, systematic, stratified and cluster sampling)
Confusion and accuracy matrices: Construction and interpretation of confusion matrices; Overall accuracy, producer's accuracy, user's accuracy, and kappa coefficient
Error analysis and improvement
Practicals
Extraction and analysis of spectral data from raster images
Analysis of metadata from raster images and its significance
Creation and interpretation of histograms for image distribution and pixel intensity
Collection and analysis of spectral signatures of materials (Vegetation, built-up, snow cover, water and soil)
Creation of composite band images with different features
Image pre-processing methods (Radiometric, atmospheric, and geometric corrections)
Application of supervised and unsupervised classification methods
Design and implementation of sampling strategies for ground truth data collection and accuracy assessments
