Satellite Imagery, Remote Sensing & Machine Learning

Coursera via Coursera Specialization

Go to class Write review

Overview

Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off

One annual plan covers every course and certificate on Coursera. 40% off for a limited time.

Get Full Access

Transform raw satellite data into actionable environmental insights with this 8-course program bridging traditional remote sensing with cutting-edge machine learning. Start by learning fundamentals: understanding how satellites measure Earth, calculating vegetation indices, & working with LiDAR 3D point clouds. Progress to advanced techniques of spatial statistics, SAR processing for disaster response, & climate data analysis. Dive into machine learning with hands-on training in CNNs for land cover classification, transfer learning, & model interpretability using Grad-CAM. Master Google Earth Engine for large-scale environmental monitoring without complex infrastructure. Through practical projects, analyze forest health, detect flood extent, evaluate air quality, & track vegetation trends. Each course emphasizes real-world application, from creating elevation models to generating climate reports for ESG initiatives. Learn to handle diverse data types—multispectral, SAR, LiDAR, & climate datasets—while building confidence in analysis & communication. Whether monitoring deforestation, assessing disasters, or tracking climate indicators, gain skills essential for environmental consulting & sustainability reporting. Perfect for GIS professionals, environmental analysts, & data scientists entering Earth observation. By completion, you'll confidently process satellite imagery, apply machine learning to environmental challenges, & deliver insights supporting critical decisions.

Syllabus

Course 1: Start Remote Sensing
Course 2: Explore LiDAR in 3D
Course 3: Process SAR & Multispectral
Course 4: Crunch Spatial Stats
Course 5: Analyze Climate Data
Course 6: Train ML Models
Course 7: Deep Learn Imagery
Course 8: Harness Earth Engine

Courses

0 reviews

View details

To take this course you must have a prior experience and basic comfort working with data and interpreting simple charts. But no prior programming, climate‑science or ESG background required. This course starts with working confidently with climate data. In this short, hands-on course, you’ll learn how to analyze real climate datasets and translate your findings into clear, sustainability-focused insights. You’ll calculate temperature anomalies from NetCDF data, visualize decade-long trends using Google Collab and Matplotlib, and interpret patterns that matter for long-term climate analysis. Through guided activities and practical examples, you’ll move from raw data to clear charts and concise, evidence-based summaries. Designed for beginner data analysts, Analyze Climate Data builds your confidence in producing climate trend insights that are ready to include in sustainability or ESG reports.
0 reviews

View details

Spatial data is everywhere, but maps alone can be misleading. In Crunch Spatial Stats, you will move beyond visual patterns and use spatial statistics to make defensible, evidence-based conclusions from location-based data. Working with realistic air-quality examples, you will develop practical skills to test whether patterns are meaningful, estimate conditions between measurements, and explain how spatial relationships change with distance. The course emphasizes clear reasoning and interpretation, not complex mathematics, so you will confidently explain results to both technical and non-technical audiences. By the end of this course, you will be able to compute Global Moran’s I for a polygon layer, perform IDW interpolation for point observations, and interpret semivariograms to assess spatial autocorrelation. Throughout the course, you will practice skills commonly used in environmental monitoring, public health, and spatial analysis roles, focusing on understanding the assumptions and limitations behind each method. This course is designed for beginners. You will need basic familiarity with maps, tabular datasets, and simple descriptive statistics. No prior experience with spatial statistics or geostatistical modeling is required.
0 reviews

View details

This hands-on course proves that deep learning isn't just about pressing "run" on a model. It's about turning satellite imagery into actual, useful insights. You'll work with convolutional neural networks for land cover classification, fine-tune a pre-trained CNN using transfer learning, use data augmentation to improve performance, and apply Grad-CAM to see where the model is actually looking. Along the way, you'll practice translating raw satellite imagery into insights you can clearly communicate to others. You are required to have basic Python programming, familiarity with machine learning concepts, and introductory knowledge of neural networks and image data. Designed for beginners in machine learning and remote sensing, Deep Learn Imagery builds your confidence in both working with deep learning and explaining what your models are doing.
0 reviews

View details

Explore LiDAR in 3D is a beginner-level course designed for analysts and GIS professionals who want to work confidently with LiDAR point cloud data for elevation modeling and environmental applications. LiDAR data is inherently three-dimensional, yet many errors in analysis occur when users treat it like traditional 2D raster data. This course focuses on building correct mental models and practical skills for working with LiDAR from visualization through validation. Across three structured lessons, learners progress from understanding and exploring LiDAR point clouds in a 3D viewer, to generating a Digital Elevation Model (DEM) from ground-class points, and finally to evaluating vertical accuracy using control points. Rather than emphasizing software clicks, the course emphasizes why each step matters, helping learners make sound decisions that support real-world use cases such as flood modelling. Through short videos, targeted readings, Coach-led reflections, and hands-on labs, learners practice loading and interpreting point clouds, creating terrain surfaces correctly, and assessing whether elevation outputs are fit for purpose. By the end of the course, learners will be able to confidently visualize LiDAR data in 3D, generate a DEM from ground points, and evaluate its vertical accuracy—skills essential for delivering reliable elevation models in environmental and risk-analysis workflows.
0 reviews

View details

Harness Earth Engine is a beginner-level course designed for learners who want to explore environmental change using satellite data without managing complex data pipelines. Across a series of focused, hands-on lessons, this course introduces Google Earth Engine as a practical tool for analyzing vegetation patterns at scale. Learners begin by understanding NDVI and its role in environmental monitoring, then progress to working directly with the MODIS NDVI dataset inside Earth Engine. Through guided videos, readings, Coach dialogues, and hands-on practice, learners build skills to load satellite image collections, reduce time-series data into annual summaries, and visualize long-term vegetation trends using charts. The course emphasizes real-world application, preparing learners to create NDVI trend charts suitable for environmental or sustainability reports. By the end of the course, learners will have a clear, repeatable workflow for transforming raw satellite imagery into meaningful, report-ready insights.
0 reviews

View details

Process SAR & Multispectral is a short course for learners who want to move beyond viewing satellite imagery and begin producing structured geospatial analysis. Designed for those with basic familiarity with maps and raster imagery, the course introduces practical techniques for interpreting and analyzing satellite data in a disaster-response scenario: estimating flood extent after a major storm. You will first work with Synthetic Aperture Radar (SAR), learning why it is essential when clouds block optical imagery and how speckle filtering can improve interpretability while introducing analytical trade-offs. The course then transitions to multispectral imagery, where you explore change detection across time to identify areas where surface conditions may have shifted after the storm. Finally, you will evaluate whether your results are reliable enough to share by interpreting simple accuracy metrics and identifying limitations in your analysis. Through guided videos, applied exercises, and scenario-based assessments, you will build both technical understanding and analytical judgment—preparing you for more advanced geospatial analysis workflows.
0 reviews

View details

Start Remote Sensing is a beginner-level course that introduces you to the core concepts needed to analyze vegetation using satellite imagery, with a strong focus on understanding data before drawing conclusions. In this course, you explore how Earth-observing satellites collect measurements not photographs and why those measurements must be interpreted carefully. You will compare commonly used satellite missions such as Landsat and Sentinel, learning how differences in spatial resolution, revisit frequency, and spectral bands influence what can be analyzed and how confidently results can be compared. You will then calculate and interpret the Normalized Difference Vegetation Index (NDVI) using red and near-infrared bands, focusing on what the index responds to and what its values reveal about vegetation condition. The course also emphasizes data readiness, showing how atmospheric effects distort raw satellite imagery and why preprocessing steps such as atmospheric correction and surface reflectance are essential before NDVI analysis. By the end of the course, you will be able to produce and interpret a vegetation index that supports a forest-health brief while developing reasoning skills for further study or applied work in remote sensing and environmental monitoring.
0 reviews

View details

This course equips learners with practical, job-ready skills to train and evaluate supervised machine learning models for land-cover classification. Learners progress through an end-to-end analytical workflow, beginning with spectral and texture feature engineering, followed by training a Random Forest classifier, and concluding with rigorous validation using confusion-matrix-based accuracy assessment. By the end of the course, learners produce a land-cover map that meets a minimum accuracy threshold, mirroring real-world data analysis workflows.