GPS Surveying & Drone Mapping for Public Works and Cadaster Departments

 Background

Historically local governments in many countries used optical survey instruments to map their cadastre and public works projects. Many municipalities establish a baseline using a survey control benchmark in the square near their municipal offices to establish a baseline to measure their surveys from. 

In the Philippines the U.S. Army Corps of Engineers originally established the geographic coordinates of these control monuments using celestial observations. These coordinates have been updated with the launching of the GPS satellites by the U.S. Department of Defence in 1978. 

Since then Europe, China, Russia and others have added their own constellations of satellites to augment positional accuracy and to decrease the reliance on one nation for that service. They are referred to the Global Navigation Satellite System (GNSS) as a whole for precise geo-spatial positioning, navigation and timing (PNT) services. They are named respectively Galileo, BeiDou, Glonass respectively. GNSS receivers which have replaced the older GPS receivers can now use over 100 satellites for positioning accuracies instead of just 24 for GPS.

When GPS was originally introduced the U.S. Military intentionally degraded the signals for civilians with the introduction of Selective Availability (SA) which degraded the signals to the receivers so that its position is reduced in accuracy to 100 m until May 2000 when it was discontinued by President Clinton to enhance civilian, commercial and safety applications.

Use of Handheld GPS receivers for Survey and Mapping

Many municipalities are using hand held GPS receivers for surveying and mapping instead of total stations. It is important to understand that these hand held units typically provide accuracy to within 3 to 10 m under normal, open-sky conditions. Under open skies, high-quality handhelds can reach 1 to 3 m with multi-GNSS support such as GPS, Galileo and GLONASS.

Users must be aware of environmental obstacles such as tree cover, deep canyons, or tall buildings which causes the GPS signals to bounce, reducing accuracies to 10 m or more. There is also the phenomenon of GPS drift movements which can be up to 100 m in particularly poor conditions.

Typical Accuracy Ranges

• Standard handheld or smartphones: 3 to 10 m
• High end/multi-GNSS handheld: 1 to 3 m
• Sub-meter/Differential GPS: < 1 m

Examples of handheld receivers available to the technical staff at Copan Ruinas. The Android GPS Altitude app is an app that any staff member can download on their mobile phones for use in field data collection.

Waypoints and Tracks

Garmin receivers and mobile phones provide the capability of storing GPS coordinate data as waypoints and tracks. Waypoints are single point location in latitude and longitude in WGS84 stored in the devices. Tracks are connected waypoints describing the path of your trip. The raw coordinate data is always stored as latitude and longitude in WGS84 which is also know as Earth-Centered, Earth-Fixed datum that the satellites rotate around. 

Users may display the waypoints and tracks in a projected coordinate such as WGS84 UTM in their specific Zone, e.g., Copan, Honduras is in UTM Zone 16 N on the display but the raw stored data is in degrees and the result is transformed to display the units in a flat plane for mapping purposes. The raw data in Garmin receivers are stored in GPX format.

Google Map Images and Open Street Map (OSM)

Google earth used Landsat imagery from its mapping program for satellite images around the world. It supplemented the images with newer, more detailed images as they became available from others that have requested the imagery. Landsat mapping program started in 1972 and continues to date. This imagery continues to be used for large scale remote sensing for environmental projects. However the limitation of is the size of the pixels were 15 to 30 m not accurate enough for detailed needs.

If the newer satellite imagery has not been rectified (adjusted with accurate ground control points (GCP)), and an algorithm is used to position it on top of older imagery it may have position errors with respect to waypoints and tracks collected. Some municipalities have adjusted their track data to fit the Google Earth maps instead of the other way around. If your waypoints and tracks are consistently different than the GE image you may want to move the image to fit the GPS data instead of the other way round.

Google maps uses the Web Mercator projection (EPSG:3857 or Psuedo-Mercator) for its 2D map views. This is a variant of the spherical Mercator projection designed specifically for web mapping, because it keeps angles, shapes and street directions accurate for navigation, even though it distorts the size of areas far from the equator. 

How to improve accuracy with your Garmin and your single phase GPS receivers

1. Use the waypoint averaging option. Accessible via the Waypoint Manager menu. This feature is best used by collecting 4 to 8 samples, ideally 90 minutes apart, until the confidence bar reaches 100%.
2. Turn on Wide Area Augmentation System (WAAS), which is a satellite based navigation aid that enhances GPS accuracy. It uses ground stations and geostationary satellites to improve position accuracy to less than a meter depending on your GPS receiver. It was designed to help aircraft make a Category I approach at airports that are equiped for it.
3. If your GPS unit supports storing of data in Receiver Independent Exchange Format (RINEX) standard ASCII format you can use it to establish your static benchmark with post processing using Precise Point Positioning (PPP) by uploading your RINEX observations to the nearest GNSS reference receiver station. These stations enable single-receiver, centimeter-level accuracy worldwide. Alternatively you could use local, real-time base stations that offer NTRIP casting for centimeter-level positioning accuracy by streaming RTCM-formatted data over the internet such as the EMLID Caster.
4. With well established benchmarks you could use GPS receivers that allow you to apply differential survey points with one GPS receiver on a benchmark (base station) and the other moving from point to point (rover). This could be achieved with the EMLID Reach RS receivers.

Map Projections & Datums

Map projections are the mathematical transformations used to represent the 3D, curved surface of the Earth on flat 2D maps and plans. Inevitably causing distortion in shape, area, distance or direction. Common types include cylindrical, conic and azimuthal, selected based on whether the goal is to preserve angles (conformal), area (equal-area) or distances (equidistant).

Geographic Coordinates - Latitude & Longitude

The earth is a sphere and a position can be determined by a unique pair of coordinates in degrees northern hemisphere (north of the equator). The latitute is zero degrees at the equator and latitude is measured from the Prime Meridian (Greenwich) east and west of it. Longitude is measured in degrees north starting at 0 degrees on the equator to 90 degrees at the North Pole. 

This coordinate system has been used by sailors since the time of the Greeks in CE 130. 

Universal Transverse Mercator (UTM)

The linear distances and areas in degrees varies around the world so it cannot be used for local measurements. Consequently projections have been used to flatten the 3D sphere to a 2D flat map. The most common projection we use today to represent flat 2D maps is the WGS84 UTM projection which has divide the world in 6 degree strips starting from the Prime Meridian and numbered as Zones. For example Honduras is in UTM Zone 16N.

UTM Zone 16N  is a 6 degree zone between latitude 84 deg. W and 90 deg. W of the Prime Meridian. Its Central Meridian is 87 deg W. The UTM north axis is 0,0 at 87 deg. W and 0 deg N (of the Equator). To prevent negative numbers in the X-axis a false Easting of 500,000 m is added to 0,0. So that any X value within the zone will never be negative.

Map Datums

A map datum is a mathematical model of the Earth's shape, providing a reference framework (origin, orientation and elevations) for measuring postions, latitudes, longitudes and elevations. It ensures that coordinates (e.g., GPS) align accurately with surface features. 

Common types of datum include geocentric (WGS84, World Geodetic System 1984) and local/regional datum like NAD27 and NAD83. NAD27 uses the Clarke 1866 ellipsoid and NAD83 uses the GRS80 ellipsoid which is similar to the WGS84 ellipsoid which are Earth Centered Earth Fixed (ECEF).

Australian Geodetic Datum (AGD84) left and Geocentric Datum of Australia (GDA94) right.

NAD27 and NAD83 which is an update for North America could be 30 m plus difference in their coordinates. NAD83 and WGS84 are nearly identical except that there is more flattening of the ellipsoid with GRS80 than WGS84.

It is recommended that Copan uses UTM WGS84 or NAD83. The main differences between the two is the elevations not the horizontal position. 

Copan Ruinas Example Waypoint Averaging

Processing Garmin Waypoint and Tracks

The Minnesota Department of Natural Resources developed two applications for users of Garmin GPS receivers. They were DNRGARMIN and the newer DNRGPS. Both work fine to download data from Garmin units.

Garmin GPS receivers stores the track data in the GPX folder for example if you connect to the receiver with your USB cable you could get: D:\Garmin GPSMAP 64s\Garmin\Current. Your saved waypoints are stored in the \Garmin folder by date, all waypoint files for the day is stored together in one waypoint gpx file. So it is important to use the comment option for averaged values or at least log the point number and averging details.

Your track data is stored all together in the \Current folder in the Current.gpx file. You can store 10,000 track points after which, if you selected the option new values will replace the oldest point values. So it is important to archive and clear the Current.gpx file regularly. You can save the waypoint and current gpx file to your PC before processing or access it directly on your Garmin receiver. We recommend you download the data and open it from your harddrive. It has difficulties sometimes to connect via USB.

You should use DNRGPS to save the waypoint and track files to QGIS shape files because there seems to be a problem importing GPX files into GIS. Note that the raw GPX coordinates are in WGS84 latitude and longitude and you can display projected coordinates in the table as well.

Waypoint Averaging and Track points

We have used the town hall benchmark in the square to test GPS waypoint averaging with the Garmin receivers. The units were also setup to collect Track data continuously every 5 seconds. The results are shown below:

Two averaged points are shown in yellow they are less than 3 m apart. The magenta points are the Track points collected every 5 seconds. At any given time they could be off by 10 m or more. Consequently field points should be collected by averaging. Benchmarks should be averaged more frequently and over a longer period like 30 to 60 minutes if you want a better benchmark coordinate.

GPS/GNSS with Mobile Apps

Surveying pipelines and intake locations are very difficult under the tree canopy. Total stations require line-of-signt and GPS/GNSS receivers require clear sky overhead to locate the satellites. For total station surveys trees and bushes need to be cleared to log observations. For GPS/GNSS receivers a large clearing is needed to log a good waypoint position.

Most new mobile phones now have GNSS receivers built in which allows postions to be collected using GPS, Galileo, GLONASS, etc. constelations for position triangulation. In some ways the mobile phone seems to be able to log positions more readily than dedicated hand held units. Because they are receiving GNSS satellites they may even provide better positional accuracy.

It is my opinion that it is possible to use mobile GPS/GNSS apps waypoints and tracks for conceptual and preliminary engineer design, given an understanding of the accuracies we are working with.

GPS Altitude

This is a free Android app which can be used to collect waypoints and altitudes under tree cover. However it needs to be noted that the accuracies can vary based on conditions. 

Note: the free version comes with adverstisements that you have to navigate around. You can wait until the Coordinate accuracy decreases to 2 m before saving the point, if that is possible. Saved points can be shared with your computer.

Geo Tracker

Geo Tracker was another Android app that was used to track the routes to the water intakes in Copan Ruinas. It appears to have worked quite well under tree cover. It should be noted that the accuracies may vary due to the conditions. However it is a good tool to map the potential pipeline routing under tree cover. It could be followed up with a more detailed survey if required.

Geo Tracker is a good GPS/GNSS app that captures tracks. This app can be used to map a road network in your community, it can also be used for conceptual and preliminary engineering layouts for construction projects.

GPS/GNSS Elevations

Most users of GPS/GNSS data do not understand the elevation displayed with their waypoint location. Google Earth displays elevations at the location of your cursor. This elevation is based on the Shuttle Radar Topography Mission (SRTM) to collect Digital Terrain Elevation Data (DTED) data. It has a minimum vertical accuracy of 16 m absolute error at 90% confidence (Root Mean Square Error (RMSE) of 9.73 m) world wide. 

SRTM used WGS84 ellipsoid for its horizontal coordinates and the Earth Gravitational Model 1996 (EGM96) geoid for its vertical datum. Modern professinal high accuracy GPS/GNSS receivers may use more refined models like EGM 2008. This is the same geoid that is used by your Garmin receivers for their estimate of elevations.

It should also be noted that SRTM  used a C-band radar which did not always map the absolute ground level. In forested areas, the elevation data often represented the top of canopy (tree heights) rather than bare ground.

Emlid Reach RS

The Emlid Reach RS is a single phase receiver that is able to save RINEX data. A set has been made available to the town of Copan Ruinas for their surveying needs. This allows their staff to establish benchmarks with PPP and it allows them to undertake differential point surveys with one unit as a base station on a known benchmark and the other as a rover collecting points selectively in the project area.

An example of archeologists surveying in Porto Rafti, Greece is shown below.

Drone Mapping (Photogrammetry)

Background

This component of the blog could be a stand alone section, however the preceding information on GPS surveying and the underlying principles of satellite positioning is an important component in drone mapping. Hence I have combined the two. Over 10 years ago I started working with drones to collect point elevations for landfill sites with a surveyor colleague. 

The reason to use a drone is simple. It would take less time to fly and map a site than to survey it with a total station. It would be possible to get a denser point file in much less time. It would also be possible to undertake the exercise with just one person. However, it would require that ground control points (GCP) be established to register the air photo.

The first drone we used was a 3D Robotics IRIS which allowed us to mount a camera (e.g., GoPro) that could be programmed to take photos at specified frequency. It supported mission planning which is needed to instruct the drone to take photos of the ground at specific waypoints in order to be able to process them into a photo mosaic map of the site. The initial application for Windows/Linux was Mission Planner.

In 2015 DJI introduced the Phantom 1 which included an integrated camera and gimbal. In 2016 I used the Phantom 3 for drone mapping in Mongolia as part of my assignment to train government staff involved in environmental management. Drone Deploy was used for mission planning at that time. A blog of training in Mongolia:  MERIT Drone Training in Sukhbaatar Aimag 2018

Drone Mapping in 2026

Over 10 years ago we found that drones could be used to map and contour smaller project areas in a cost-effective way. That industry and technology has come a long way. Unfortunately that has also resulted in more restrictions and consolidation has resulted in cost increases. So can we still use drones for mapping if we have limited budgets? Well, my answer is yes.

Today most governments requires that drones over 250 g be registered and pilots of those drones have to be certified. However drones under 250 g are considered to be mini drones good for recreational purposes and so they do not require registration. They are also under USD 1,000. 

Unfortunately DJI has decided that it will not provide 3rd party developers of drone apps the software development kit (SDK) that is needed for them to provide waypoint and mission planning apps for these drones. However I found one exception for the DJI mini 2 and 2 SE. Litchi provides an Android app that will work with it. 

There is also a work around for the DJI mini 4 Pro which will allow us to load a mission plan to the drone using fly litchi hub website and the DJI Fly app. There are some steps to follow but overall it works. It is good that Litchi has provided help for this work around on their website.

Flight Planning & Mission Considerations

There are some key considerations in flight planning and mission execution:

1. Flight time - the estimate battery time is 20 to 25 minutes. We recommend a mission flight time of 15 minutes to ensure safe return to home.
2. Wind direction - We recommend that the main legs of the flight be in line with the prevailing wind direction. The drone flight characterisitics would be more stable.
3. Wind force - We also recommend that you fly during calm weather. Windy conditions causes the drone to be less stable and affect the quality of the photographs.
4. Time of day - It is best to fly when the sun is higher in the sky so there are less shadows in the photographs.
5. Weather - Obviously cloudy and rainy days would not be the best time to fly.
6. GCPs - It is always best to have good ground control points set up to register your images. You can get an orthophoto without GCPs but it will not be as accurate.

Photogrammetry

The drone camera is set to 90 degrees (pointing down) to take images of the ground perpendicularly. The mission software automatically establishes the next waypoint to take a photo with the required overlap to the last photo. These photos are overlapped in the forward direction and side overlapped when the next waypoint line is flown in the opposite direction in the next row.

The principles of photogrammetry is to use the overlapped area of the 'stereo-pairs' to determine parallax and stereo 3D for the estimation of elevations. In the past optical instruments called stereoplotters were used to create contours on maps in the overlapped areas. Today computers are used to automattically create point clouds of xyz data which can be processed into contours or DEMs.

The photos from the flight missions are loaded into an application like WebODM to create point clouds and orthophotos.

Tutorial Using Litchi Mission Hub

Step 1 - Litchi Website Launch Mission Hub

On the Litchi website (Litchi Mission Hub) select the Mission tab option to create your mission waypoint file. Use your mouse to move to the location of your project. Litchi Mission uses Google Earth type base map you can move, zoom in and out to your selected project geo-location.

Step 2 - Create Flight Waypoints

Select your first start waypoint and enter the drone flight parameters in the waypoint popup form. An example is shown in the figure below.

 

Enter the Altitude (m) value of 80 m above the 'starting elevation' of the drone. Note: you can set the drone to follow the ground by activating the Above Ground button, but you do not want that for photogrammetry. You want the drone to fly in a fixed plane above the ground, however you need to make sure that the Altitude selected will be clear of any ground obstacles. 

In our example Speed (m/s) is set at 5 m/s. We have set the Heading of the drone to follow the flight lines. You can do that by moving the slider button while looking at the direction arrow on the waypoint on the screen in Mission Hub.

The next parameter to set is the gimbal Angle (degrees). It is set at -90 degrees to point the camera on the gimbal to the ground. Zero degrees is the horizontal setting for the camera. The Interval is the time between each picture frame and it is set to 3.0 seconds.

These values are set for each waypoint on the mission plan.

To calculate image overlap, etc. you can use Sitephotos website.

Step 3 - Complete and Save the Mission File

An example of our tutorial mission is shown in the exhibit below.

The mission starts at Waypoint 1, note the elevation is set at 80 m above the drone start elevation. The first leg between Wapoint 1 and 2 is 329 m. Note the direction pointer (heading) on the waypoint; it is pointing to Waypoint 2, etc. The total mission is approximately 1.6 km and the time of the mission is estimated to be 7 minutes based on the drone speed of 5 m/s.

Once the mission file is saved to the Litchi cloud server it will be available to your mobile phone in the field when you select the Waypoint mission as the option on your Litchi app.

The results of this tutorial mission will be added to this blog once it is completed.