top of page
SKYDOG ATI LLC
Professional Drone Thermal Imaging & SAR Services in New England.
Frequently asked questions
General
GPS Accuracy Overview(https://www.skydogati.com/gps-accuracy)
RTK / PPK Specifics
Gemini 2.5 PRO: "When using an RTK-enabled drone with an RTK base station set up at an unknown location, the relative accuracy between the drone and the base station remains very high—typically at the centimeter level.
However, the absolute accuracy of the entire system (where the drone and base are located on Earth) will be low, likely off by several meters.
How It Works: Relative vs. Absolute Accuracy 🎯
Think of the RTK base station as an anchor point. In a standard setup, this anchor is placed on a precisely known coordinate. In your scenario, the anchor is dropped at an unknown location.
1. The Base Station Sets a Reference: The base station turns on and calculates its position using standard GNSS signals. This initial position will have an error of a few meters. This inaccurate point becomes the "Point A" for your entire operation.
2. Error Correction is Shared: Both the base station at "Point A" and the RTK drone flying nearby are looking at the same satellites through the same atmospheric conditions. Therefore, they experience the exact same GNSS signal errors at the same time.
3. Relative Precision is Achieved: The base station calculates these shared errors and broadcasts a correction signal to the drone. The drone uses this signal to cancel out its own errors. The result is that the drone can calculate its position with centimeter-level precision relative to the base station's "Point A".
The system doesn't know where "Point A" truly is on the globe, but it knows with extreme precision where the drone is in relation to "Point A".
Analogy: Imagine you're drawing a detailed map of a building.
• High Absolute Accuracy: You use a survey-grade GPS to place the corner of the building on the map at its exact latitude and longitude. The map is perfectly located.
• High Relative Accuracy (Your Scenario): You just start drawing the building anywhere on the page. The dimensions of the rooms, the length of the hallways, and the angles of the walls are all perfectly drawn to scale (centimeter-level relative accuracy). However, the entire drawing might be in the wrong spot on the page (meter-level absolute accuracy).
When using an RTK-enabled drone with an RTK base station set up at an unknown location, the relative accuracy between the drone and the base station remains very high—typically at the centimeter level.
However, the absolute accuracy of the entire system (where the drone and base are located on Earth) will be low, likely off by several meters.
How It Works: Relative vs. Absolute Accuracy 🎯
Think of the RTK base station as an anchor point. In a standard setup, this anchor is placed on a precisely known coordinate. In your scenario, the anchor is dropped at an unknown location.
1. The Base Station Sets a Reference: The base station turns on and calculates its position using standard GNSS signals. This initial position will have an error of a few meters. This inaccurate point becomes the "Point A" for your entire operation.
2. Error Correction is Shared: Both the base station at "Point A" and the RTK drone flying nearby are looking at the same satellites through the same atmospheric conditions. Therefore, they experience the exact same GNSS signal errors at the same time.
3. Relative Precision is Achieved: The base station calculates these shared errors and broadcasts a correction signal to the drone. The drone uses this signal to cancel out its own errors. The result is that the drone can calculate its position with centimeter-level precision relative to the base station's "Point A".
The system doesn't know where "Point A" truly is on the globe, but it knows with extreme precision where the drone is in relation to "Point A".
Analogy: Imagine you're drawing a detailed map of a building.
• High Absolute Accuracy: You use a survey-grade GPS to place the corner of the building on the map at its exact latitude and longitude. The map is perfectly located.
• High Relative Accuracy (Your Scenario): You just start drawing the building anywhere on the page. The dimensions of the rooms, the length of the hallways, and the angles of the walls are all perfectly drawn to scale (centimeter-level relative accuracy). However, the entire drawing might be in the wrong spot on the page (meter-level absolute accuracy).
Summary of Accuracy
Accuracy TypePrecision LevelExplanationRelative Accuracy (Drone to Base)Centimeter-levelThe distance and vector between the base and the drone are known with very high precision. All data collected by the drone is internally consistent.Absolute Accuracy (Drone to the Earth)Meter-levelThe entire dataset (the base and the drone's flight path) is shifted by the same amount as the base station's initial position error.
Export to Sheets
Practical Implications & Use Cases
This method is highly effective for jobs where the internal consistency of the data is more important than its true global position.
✅ Good Use Cases:
• Volume Calculations: Measuring the volume of a stockpile (e.g., gravel, sand). The exact location of the pile doesn't matter, only its precise shape and size.
• 3D Modeling: Creating a highly detailed but unreferenced 3D model of a building, accident scene, or small area.
• Relative Change Detection: Monitoring erosion or movement on a small scale, where you only care about how much things have changed relative to the last flight.
❌ Bad Use Cases:
• Surveying & Boundary Work: You cannot use this data to establish property lines or for any job requiring legally defensible geodetic coordinates.
• GIS Overlays: Your map or model will not align correctly with other geographically accurate data layers (e.g., utility maps, property boundaries, Google Earth imagery)."
The absolute best resolution you will find on Google Maps and Google Earth is 15 centimeters (about 6 inches) per pixel (via airplanes over cities). At this level of detail, you can clearly make out individual car models, the painted lines in a parking lot, and even the shadows cast by people.
However, there are three distinct tiers of resolution on the platform, and the crispest images aren't actually taken from space:
• The 15 cm Peak (Aerial Photography): The ultra-crisp imagery you see in major urban areas (like New York, London, or Las Vegas) is captured by specially equipped airplanes flying relatively close to the ground, not satellites.
• 30 to 50 cm (True Satellite Maximum): For areas where planes don't fly, Google relies on commercial satellite providers like Maxar. The sharpest actual satellite imagery currently available on the platform maxes out at 30 centimeters (about 12 inches) per pixel.
• 15 Meters (The Global Baseline): If you scroll over oceans, remote wilderness, or less populated global regions, the resolution drops to about 15 meters per pixel. This baseline layer is largely stitched together from public satellite programs like NASA's Landsat or the ESA's Sentinel.
Sky Dog ATI typically deliver accuracy of 1 - 2 cm per pixel while flying from typically 150 - 200' AGL.
bottom of page