Definition: What is GSD?
Ground Sampling Distance (GSD) is the size of a single pixel on the ground, measured in centimeters. If your drone flies with a GSD of 2.74 cm, it means: each pixel in the image corresponds to a square of 2.74 × 2.74 cm on the ground – or 7.52 cm² of area.
GSD is the key to image resolution. It determines not only how detailed your 3D model becomes, but also the achievable measurement accuracy. A smaller GSD (e.g. 0.5 cm) produces sharper details and higher accuracy, but requires a lower flight altitude or better camera sensors.
In practice, you work with GSD value ranges: A GSD of 1–3 cm is completely sufficient for most surveying tasks (photogrammetry, orthophotos, roof property capture). For specialized applications like plant identification, GSD values below 0.5 cm are required.
GSD is similar to the DPI (Dots Per Inch) concept from the printing industry. With drone imagery, we speak of cm per pixel instead of inches.
GSD Memory Rule
Smaller GSD = higher resolution = more detail. But: this requires either a lower flight altitude OR a better camera. The formula shows the exact relationship.
The GSD Calculation Formula
There are several equivalent representations of the GSD formula. The most common version is:
GSD = (Sensor Width × Flight Height) / (Image Width × Focal Length)
Where the parameters are:
• Sensor width (in mm): The physical width of the drone's image sensor – NOT the resolution in megapixels, but the actual sensor dimensions (e.g. 13.2 mm for Mavic 3)
• Flight height (in m): The height at which you fly – measured from the ground to the drone
• Image width (in pixels): The horizontal pixel resolution of the image (e.g. 5472 pixels)
• Focal length (in mm): The physical focal length of the lens – NOT the 35mm equivalent (e.g. 8.8 mm for Mavic 3, not 24 mm)
The result is the GSD in meters per pixel. To convert to centimeters, multiply the result by 100.
Example: With DJI Mavic 3 with sensor width 13.2 mm, image width 5472 px, focal length 8.8 mm and flight height 100 m:
GSD = (13.2 × 100) / (5472 × 8.8) = 1320 / 48,153.6 = 0.02740 m/px = 2.74 cm/px
Common Mistake: 35mm Equivalent vs. Actual Focal Length
DJI lists 24 mm for the Mavic 3 – but that's the 35mm equivalent. The actual physical focal length is 8.8 mm. ALWAYS use the physical focal length in the formula, not the 35mm equivalent. Using the wrong number will make your GSD calculation wrong by a factor of 2.7!
Influencing Factors: Sensor, Focal Length, Altitude
GSD is determined by exactly three factors:
1. Sensor size (sensor width/sensor height) – determines the physical size of each pixel
2. Focal length of the lens – longer focal length = smaller GSD (better resolution)
3. Flight altitude – higher flying = larger GSD (worse resolution)
The formula shows: GSD is directly proportional to flight altitude and inversely proportional to focal length. This means concretely:
• Double the flight altitude from 50 m to 100 m → GSD doubles (2× worse resolution)
• Double the focal length (longer lens) → GSD halves (2× better resolution)
• With a larger sensor (more mm sensor width) → GSD gets larger (worse resolution) – this is why full-frame cameras in drones are disadvantageous
For practitioners, this means: To maintain a certain GSD as flight altitude increases, you need a longer focal length or a smaller sensor. This is why professional surveying drones often use zoom cameras with variable focal lengths.
Choose Flight Altitude Wisely
Choose the lowest flight altitude your mission allows (obstacles, flight safety, regulations)! This saves you from the most expensive option: upgrading your camera setup. Most surveys can be accomplished with flight altitudes of 50–120 m – that is the standard.
Practical GSD Values for Popular Drones
To avoid recalculating the formula every time: here are the GSD values for the most-used surveying drones at standard flight heights and standard cameras:
GSD & Accuracy: The Relationship
A common misconception: GSD = absolute accuracy. That's not correct. GSD is more of an upper limit for resolution, not accuracy itself.
The rule from technical literature and practical experience is: The horizontal accuracy of a well-reconstructed photogrammetry is between 1x and 3x the GSD. This means:
• With GSD = 2 cm you can expect realistic horizontal accuracies of 2–6 cm.
• With GSD = 1 cm you achieve 1–3 cm horizontal accuracy.
• With GSD = 0.5 cm, 0.5–1.5 cm is possible.
The exact position within this range depends on:
• Quality of image overlap (at least 60% side overlap, 70% longitudinal overlap)
• Camera calibration and lens distortions
• Use of ground control points (GCP) or RTK-GNSS – RTK provides better absolute accuracy
• Terrain and lighting conditions during flight
Vertical accuracy (elevation) is typically worse than horizontal accuracy – one rule says: vertical accuracy ≈ (1–3) × horizontal accuracy. With good RTK solution, vertical accuracy can be better than GSD.
RTK Significantly Improves Absolute Accuracy
RTK-GNSS (Real Time Kinematic) does not improve GSD, but does improve absolute georeferencing. With RTK you typically get ±2–3 cm absolute accuracy. Without RTK (only drone barometer) it can be ±10–20 cm error. GCP (ground control points) is a cheaper alternative.
Use Cases: When Do You Need Which GSD?
Different applications require different GSD values. Here's an overview:
"Very coarse overview photography, event documentation" – 5–10 cm GSD: More suitable for video than for measurement.
"Roof inventory, building perspectives, overview models" – 2–3 cm GSD: Sufficient to identify building structures and create simple 3D models.
"Photogrammetric 3D models, PV roof planning, facade surveying" – 0.5–1.5 cm GSD: The standard for professional surveys. With this resolution you get correct roof dimensions for PV planning.
"High-precision surveying, construction site as-built comparison per VOB" – 0.3–0.5 cm GSD: This requirement requires special cameras or very low flight altitudes.
"Agriculture, plant health, single plant recognition" – < 0.1 cm GSD: Only achievable with ground-based photogrammetry or special multispectral drones.
For the typical Voxelia application (3D roof models for PV, facade documentation, as-built plans) the optimal GSD is 1–2 cm. This delivers outstanding results without unnecessary extra altitude and associated flight safety concerns.
Order drone survey
High-resolution orthophotos and elevation models
Voxelia delivers drone surveys with optimized GSD for your requirements – from overview shots to high-precision surveys.
Get quotePractice Tips: Plan Flight Height Correctly
"Plan backwards": You know which GSD you need, but not how high you should fly? Then solve the formula for flight height:
Flight Height = (GSD × Image Width × Focal Length) / Sensor Width
Example: You want to achieve 1.5 cm GSD with a Mavic 3:
Flight Height = (0.015 m × 5472 × 8.8) / 0.0132 = 54.9 m
So you would need to fly at approximately 55 m altitude. This is often impractical if obstacles are closer. Then you either need to:
• Live with larger GSD (e.g. 2.5 cm at 100 m altitude)
• Conduct multiple flights at different altitudes (e.g. 50 m for details + 100 m for overview)
• Switch to a drone with longer focal length / better sensor (e.g. Air 3 instead of Mini 4 Pro)
For customer projects, I recommend:
• Base GSD target: 1.5–2 cm (cost-effective and delivers good results)
• Standard flight altitude: 100 m above ground (safe, efficient, usually no legal issues)
• Difficult terrain: 50 m above ground for better details; plan multiple flights
• Use GCP/RTK: Helps with poor GPS reception and significantly increases absolute accuracy
Use Drone Planning Software
Most mission planning apps (DroneDeploy, Pix4Dcapture, openDroneMap) have GSD calculators built in. You can directly input your desired GSD, and the app automatically calculates the flight altitude. This saves time and calculation errors.
Frequently Asked Questions
Frequently Asked Questions about GSD
Article Tags
