R-Pod, which stands for Rapid, Rich and Reliable Photogrammetry On Demand.

This is a project initiated by HEIG-VD - the University of Applied Sciences of Western Switzerland, based in Yverdon-les-Bains - and led by Nicolas Delley.

Learn more about R-Pod here.

Project description


  • National Parc of Banco, Abidjan, Ivory Coast
  • Forest of Tanoe and palm plantations, Nouamou, Ivory Coast (50 km from Ghanian border)
  • Reserve of Lamto (Savanna) and Adiopodoumé forest, Ivory Coast

Challenge: Evaluate the feasibility of dense forest cartography (botanic) and its limits; create cartography of non-forestry zones.

Difficulties: In dense, humid forest the single flight coverage of the swinglet CAM was limited to 1 - 2 sq. km (0.4 - 0.8 sq. mi.) due to the high image overlap and image resolution required, and the very limited take-off and landing locations.

Flights: 10 flights to cover multiple areas.

Altitude: 270 - 400 m (183 - 1,312 ft.) above ground.

Precision: GSDs of 10 cm / 4 in. per pixel (Reserve of Lamto) and 15 cm / 6 in. per pixel (dense forest).

Experience and results

1. National Parc of Banco, Abidjan, Ivory Coast

The R-Pod tean was able to document the extent and dynamics of illegal human occupations by garage owners in the North of the National Park.

The occupied zone runs along an electrical power line, which is effectively cutting the forest in half, and was measured to be about 100 m wide.

Triangle of Sangbé, progression of human occupation in the North of the National Park of Banco

2. Forest of Tanoe and palm plantations, Nouamou

The Forest and palm plantations were mapped at standard parameters of 80% longitudinal and 60% lateral image overlap (flight altitude over ground: 300 m) allowing automated assembly of the images with the drone's Postflight image-processing software.

The resulting orthomosaic documents the extent of these plantations and allowed R-Pod's team to determine the percentage of forest left in this zone. This orthophoto also promises to be interesting for botanists, allowing them to identify tree varieties whereas otherwise this bog forest is very difficult to access.

Bog forest and oil palm cultures near Nouamou

3. Reserve of Lamto

The images collected allowed the R-Pod team to generate an orthomosaic with a ground sampling distance of 10 cm (3.9 in).

This orthomosaic was used to classify and count plant species. Further topological differentiation of the diverse Savanna was otherwise difficult at this time of year (November) since all Savanna were in full vegetation. However the drone-sourced images were still of great interest for documenting the exact area occupied by forest.

A section of the Reserve of Lamto, showing a French research center

4. Adiopodoumé forest

Mapping dense forest was found to require a flight altitude of around 400 m (1,312 ft.), and 90% longitudinal & 75% lateral overlap.

If this flight altitude was reduced to 270 m (886 ft.) the lateral overlap needed increasing to 80% in order to still allow the automatic processing of the images.

Furthermore, the R-Pod team found that an overly detailed digital elevation model of forested areas can disturb the projection of tree images and produces distorted crowns. Therefore a simplified DEM is required for this specific projection, in order to allow botanists to interpret the results.

Phenomena of tree "timidity"