PPK – Post-Processing Kinematic for Drone Surveying

What is PPK? Definition and Distinction from RTK

Post-Processing Kinematic (PPK) is a GNSS processing method where a drone records raw satellite measurement data during flight – so-called carrier phase observations – and this data is only evaluated after the flight together with simultaneous recordings from a base station. The result is centimeter-accurate positions for every captured image point.

PPK is the direct sibling of RTK (Real Time Kinematic). The key difference: with RTK, base station correction data is transmitted in real time via radio or mobile network to the drone and processed immediately. With PPK, the raw data from both receivers (drone + base station) is loaded into evaluation software on the ground after the flight. This makes PPK independent of radio connections and mobile network coverage.

In drone photogrammetry, PPK serves the same purpose as RTK: it delivers highly accurate, georeferenced camera positions that serve as control points in the photogrammetry software (Agisoft Metashape, Pix4D, DJI Terra). The result is a centimeter-accurate orthophoto, elevation model or 3D model – without needing to survey ground control points in the field.

Technical Principle: GNSS Carrier Phase Measurement and Double Differencing

The core of PPK is GNSS carrier phase measurement. Normal GPS receivers measure the travel time of the code signal (pseudoranges) – accurate enough for navigation (3–10 m), but too inaccurate for surveying. High-precision receivers additionally measure the phase of the carrier wave itself. The GPS L1 signal has a wavelength of approximately 19 cm – measured to a fraction of this wavelength, this provides millimeter resolution.

The raw carrier phase observation equation is:

Φ = ρ/λ − (c · δt_r)/λ + (c · δt_s)/λ − I/λ + T/λ + N + ε

Where: ρ = geometric distance satellite–receiver, λ = carrier signal wavelength, δt_r and δt_s = clock offsets of receiver and satellite, I = ionospheric delay, T = tropospheric delay, N = integer phase ambiguity and ε = measurement noise.

Through double differencing between two receivers and two satellites simultaneously, almost all error sources cancel out. The simplified result is: ∇∆Φ = ∇∆ρ/λ + ∇∆N + ∇∆ε. Once the software resolves ∇∆N as integers (a "Fix" solution), the drone position is determined to 1–3 cm accuracy.

Quality indicator: A Fix Ratio ≥ 95% with RMS < 2 cm indicates a high-quality PPK solution.

PPK vs. RTK vs. GCP: Direct Comparison

All three methods achieve similar absolute accuracies of 1–3 cm horizontal when executed well. The difference lies in workflow, requirements and costs:

Which Drones Support PPK?

Not every drone supports PPK. A prerequisite is a GNSS receiver that can record and export raw carrier phase observations. Simple GPS chips in consumer drones only output processed positions – no PPK possible. The following systems are known and proven for PPK:

Base Stations for PPK: CORS Networks and Own Station

For PPK, besides the flying GNSS rover (drone), you need a stationary base station that records raw data throughout the entire flight. The baseline between station and drone should remain under 30 km to minimize ionospheric modeling errors.

Option 1 – Own portable base station: Devices like the Emlid Reach RS3 (approx. €1,390) or Trimble R10/R12 can be deployed as temporary base stations. The station is set up on a known or measured point, logs raw data during the flight, and the logs are processed together with the drone logs. Advantage: full control, no ongoing costs, deployable anywhere.

Option 2 – SAPOS (Germany): The satellite positioning service of German survey administrations is operated by the state surveying offices together with the Federal Agency for Cartography and Geodesy (BKG). Over 270 permanent CORS stations are available nationwide. For PPK, archived RINEX raw data from the nearest SAPOS station can be obtained via the data portal.

Option 3 – International CORS networks: In Austria, the BEV (Federal Office for Metrology and Surveying) operates the APOS network, in Switzerland swisstopo operates the swipos-NAP/NAV network. Both also offer archived RINEX data for PPK.

Practical note on baseline length: The accuracy degradation from ionospheric residuals is approximately 1 ppm (1 cm per 10 km baseline). At a 10 km baseline, an error component of < 1 cm is expected. Up to 30 km, total accuracy remains sufficient for most surveying tasks.

PPK Workflow Step by Step

A complete PPK workflow consists of seven phases, from mission preparation to integration of the corrected positions into photogrammetry software:

1. 01

   Pre-Flight Preparation

   Activate GNSS recording on drone (Raw-Log mode). Set up own base station or identify CORS station. Ensure base station and drone use the same GNSS systems and frequencies. Wait for at least 10 satellites with PDOP < 2.

2. 02

   Start Synchronous Logging

   Base station must start recording BEFORE the first drone takeoff and end AFTER the last drone landing. Time overlap is critical – log at least 5 minutes before and after the flight for ambiguity initialization.

3. 03

   Drone Flight

   Execute photo mission as planned. Drone automatically records GNSS raw data in the background. Ensure the drone has no GNSS signal interruptions from buildings – these prevent ambiguity resolution.

4. 04

   Collect Logs

   After flight: rover log from the drone (UBX or RINEX format). Base log from own station or RINEX download from CORS. Ensure both logs fully cover the same time period.

5. 05

   PPK Processing in Software

   Import rover log and base log in Emlid Studio, define coordinate system (e.g. ETRS89/UTM Zone 32N for Germany), calculate PPK solution. Result: each drone position is output with cm accuracy and a quality flag (Fix/Float/Single). Check Fix Ratio: values under 90% are a warning signal.

6. 06

   Quality Check

   Check exported trajectory: all camera trigger points must be marked with "Fix" status. Check RMS of PPK solution: < 2 cm horizontal, < 3 cm vertical is sufficiently good.

7. 07

   Import into Photogrammetry Software

   Import corrected camera stations (CSV or geotag file) into Agisoft Metashape, Pix4Dmapper, or DJI Terra. Software processes positions as basis of bundle adjustment. Result: georeferenced orthophoto, point cloud and elevation model with ±2–3 cm absolute positional accuracy.

PPK Processing Software Overview

Processing PPK raw data is a separate step before the actual photogrammetry. The following software is proven in practice:

Emlid Studio (free): The most user-friendly option. Runs on Windows, macOS and Ubuntu. Input: RINEX or UBX files from rover and base. Output: camera position file with Fix/Float/Single flags.

RTKLIB (free, open source): Developed by Tomoji Takasu. Most powerful and flexible tool, but more complex to use. Module RTKPOST for offline PPK. Particularly suited for advanced users who need full parameter control.

Agisoft Metashape (from €179/year): Can process PPK positions directly without a separate PPK tool – ideal for a seamless workflow in one tool.

Pix4D (SaaS, from €350/month): Natively supports import of PPK camera position files. Good workflow for DJI Phantom 4 RTK and WingtraOne.

DJI Terra (Freemium): Processes PPK data for DJI drones natively. Generates orthophotos and point clouds directly from PPK data. Limited to DJI ecosystem.

When Does PPK Pay Off? The Decision Guide

The choice between PPK, RTK and GCP depends strongly on the use case. These rules of thumb help with the decision:

PPK is the best choice when: (1) The project area has no reliable 4G/LTE coverage. (2) Large areas (> 20 ha) need to be covered in one flight. (3) Multiple drones fly simultaneously – one base station serves all via PPK. (4) Flight planning shows obstacles that could briefly interrupt RTK fix.

RTK is preferable when: (1) Immediate on-site results are needed. (2) Reliable 4G coverage exists in the area. (3) Individual objects (one building, one roof) need to be documented quickly.

GCP remains useful when: (1) The highest achievable accuracy (< 1 cm) is required. (2) No PPK-capable drone is available. (3) Small project areas (< 2 ha) where GCP effort is proportionally small.

A common mistake: treating PPK as a replacement for GCP checkpoints. PPK delivers accurate camera positions – but no independent quality check of the result. For demanding projects, even with PPK, at least 3 independent check points should be surveyed and verified after photogrammetry processing.

Frequently Asked Questions