Georeferencing Scanned Maps in QGIS
**Georeferencing** is the process of aligning a scanned map or image (which initially has no real-world coordinates) to actual geographic coordinates, so it can be used correctly within a GIS. ---
📋 Prerequisites
- QGIS Desktop 3.22+ installed on your computer.
- Basic understanding of GIS data concepts.
Step-by-Step Instructions
Understand the Concept
A scanned paper map is just a picture — QGIS has no idea where on Earth it belongs. Georeferencing solves this by matching known points on the image to their real-world coordinates using **Ground Control Points (GCPs)**. ---
Gather What You Need
- The scanned map or image (JPG, PNG, TIFF) - A reference layer with known coordinates (an existing georeferenced map, satellite basemap, or known GPS points) - At least **4 well-distributed GCPs** (more is better for accuracy) ---
Open the Georeferencer Tool
1. In QGIS, go to **Layer → Georeferencer** 2. Click **Open Raster** and select your scanned map ---
Add Ground Control Points
1. Click a recognizable point on your scanned image (e.g., a road intersection, corner of a building) 2. A dialog appears asking for the point's real-world coordinates: - Enter coordinates manually, **or** - Click **From Map Canvas** and select the same location on a reference basemap 3. Repeat this for at least 4 points, spread across the image (corners work best) ---
Set Transformation Settings
1. Click **Settings → Transformation Settings** 2. Choose a **Transformation Type**: - **Linear** – for already fairly accurate images - **Polynomial 1 (Affine)** – most common for scanned maps - **Thin Plate Spline** – for heavily distorted images 3. Set the output CRS (match your project CRS) 4. Set an output file name and location ---
Run the Georeferencing
1. Click **Start Georeferencing** 2. QGIS will warp the image to fit your GCPs and save a new georeferenced raster 3. Load the output into your main QGIS project — it should now align correctly with other layers ---
Check Accuracy
- QGIS shows a **residual error** for each GCP — lower is better - If a point has a large error, double-check its coordinates or remove and re-add it - Aim for a total RMS error appropriate to your map's intended use (smaller is more accurate) ---
🧠 Quick Recap & Practice Questions
## Quick Recap
- Georeferencing gives real-world coordinates to unreferenced images
- Use at least 4 well-spread Ground Control Points (GCPs)
- Polynomial 1 (Affine) transformation works for most scanned maps
- Always check the residual/RMS error before trusting the output
---
## Practice Questions
1. Why do you need at least 4 GCPs instead of just 2?
2. What does a high residual error at a GCP indicate?
3. Which transformation type would you use for a heavily warped old map?
---
*Next Tutorial: Digitization – Creating Vector Layers from Satellite Imagery*