Remote Sensing Basics¶

Goal: Understand satellite imagery — what data is available, what spectral bands tell you, and how to do basic land cover and vegetation analysis.

What is remote sensing?¶

Remote sensing = measuring something from a distance, usually with a sensor on a satellite or drone. In GIS, it almost always means multispectral imagery — pictures taken in many wavelengths beyond visible light.

The electromagnetic spectrum¶

Sensors capture bands at different wavelengths:

UV │ Visible │ Near-Infrared │ Shortwave-IR │ Thermal │ Microwave
   │ B G R   │ NIR           │ SWIR1, SWIR2 │ TIR     │ (radar)

Different bands "see" different things:

Common satellites¶

Vegetation indices¶

Combine bands to highlight features.

NDVI (Normalized Difference Vegetation Index)¶

• Range: -1 to 1
• High (>0.5) = dense, healthy vegetation
• Low (<0.2) = bare soil, urban, water

# In ArcGIS Pro Raster Calculator
(NIR - Red) / (NIR + Red)

NDWI (Water)¶

High = water bodies.

NBR (Burn ratio)¶

Compare pre/post-fire to map burn severity.

Land cover classification¶

Goal: assign every pixel a class (water, forest, urban, ag, bare).

Supervised classification¶

1. Collect training samples for each class (you click on known examples).
2. An algorithm (Maximum Likelihood, SVM, Random Forest) learns the spectral signature of each class.
3. The model classifies every pixel.

Tool in ArcGIS Pro: Image Classification Wizard.

Unsupervised classification¶

1. Algorithm groups pixels into N clusters automatically.
2. You manually label each cluster.

Use when you don't know the categories yet.

Change detection¶

Compare two scenes from different dates:

Tool in ArcGIS Pro: Change Detection Wizard.

Where to get imagery¶

Common pitfalls¶

Practice¶

Next up¶

→ Portfolio Projects — put it all together.