Machine Learning in ArcGIS : Map Land Use Land Cover in GIS

Explore what machine learning is, its supervised, unsupervised, and reinforcement types, and how to select algorithms for remote sensing tasks like land use land cover mapping.

Learn how GIS users apply ai and ml to land use and land cover mapping, change detection, and object extraction in remote sensing via classification, clustering, and deep learning.

Learn image classification for land use and land cover in GIS, covering supervised and unsupervised approaches, with algorithms like random forest and SVM, plus accuracy assessment via confusion matrices.

Explore unsupervised and supervised image classification techniques in ArcGIS. Compare pixel-based and object-based methods, training samples, and methods like maximum likelihood, support vector machines, and random tree classifier.

Discover how to download satellite data for supervised and unsupervised land use classification in ArcGIS using the semi automatic classification plugin with USGS, NASA, and Copernicus data.

Explore unsupervised image classification in ArcGIS using a sentinel composite to reveal land use and land cover by spectral similarity, compare class counts, and guide future supervised classification.

Explore common image classification algorithms, including unsupervised clustering and supervised methods like minimum distance, maximum likelihood, Bayesian, decision trees, random forest, and support vector machines, with accuracy assessment.

Learn the workflow of supervised land use and land cover classification, defining mutually exclusive and exhaustive classes, training with spectral data, classifying pixels, and generating thematic maps with accuracy assessment.

Learn to create training data for supervised classification in ArcMap 10.6 by using the image classification toolbar and training sample manager to label cropland, water, trees, bare soil, and settlements.

Perform supervised image classification in ArcGIS with a random trees classifier, training on shapefile data and applying the model to an image, then tune trees and depth and assess misclassifications.

Apply a support vector machine classifier in ArcGIS to perform supervised land use and land cover classification, compare results with random forests, and discuss misclassifications and evaluation steps.

Learn how accuracy assessment tests supervised classifications using confusion matrices and reference data, explaining overall, user, and producer accuracies with cross tabulations.

Perform accuracy assessment in ArcGIS 10.6 using reference data and ground truth for validation. Create assessment points, extract values, and compute confusion metrics to compare random forest and SVM classifications.

Explore image segmentation in GIS and remote sensing using raster data from satellites, aircraft, and drones, balancing spectral similarity and spatial detail to improve land use classification.

Acquire remote sensing imagery from OpenAerialMap for a deep learning object-detection project in ArcGIS, focusing on palm trees on Tonga Island's plantation area.

Learn to perform image segmentation in ArcGIS using the segment mean shift algorithm, adjust spectral and spatial details, and create a segmented layer for object-based analysis.

Explore object based image classification (OBIA) and pixel based approaches, highlighting a two-step process of segmentation and object based classification that combines spectral, spatial, texture, and context features.

Create training data for object-based image classification in ArcGIS by labeling segmented raster segments in the training sample manager, collecting 15–20 samples per class, and saving for classifier building.

Apply object-based image classification with UAV data and SVM in ArcGIS. Build training data and a signature file; run the classifier and generate a final land-use image.

Apply object-based image analysis with support vector machines in ArcGIS. Classify white and red roofs, create land use land cover maps, and estimate house counts and area.

Join this bonus lecture to explore remote sensing, data science, and machine learning resources for map land use land cover in GIS, and follow Geo World on YouTube and Twitter.