Synthesis on the agricultural UAV-based remote sensing systems conducted by the International Potato Center (CIP) in close collaboration with University of Nairobi and University of Missouri, and through a community of practice.

1. Agriculture case study: Drones for
agriculture in East Africa
Dieudonné Harahagazwe (CIP), Elijah Cheruiyot (CIP) and
Arnold Bett (UoN)
Drones East Africa Conference - IQPC
Nairobi, 20-21 June 2017

2. Presentation outline
1) Evolution of low-cost platforms at CIP
2) Open source sensors and software by CIP
3) Examples of drone applications in the
agricultural sector at CIP
4) Role of community of practice in generating
and sustaining drone-based technologies
5) Field work for improving crop statistics in
Tanzania – a proof of concept
6) Key messages

3. 2004
2008
2009
2012
1. Evolution of low-cost platforms at CIP
Balloons:
• Hot air.
• Hellium.
Model planes:
• Combustion.
• Electric.
Helicopter:
•Combustion.
•Eléctric.
Multirotor:
• Quadracopter.
• Octacopter.

4. 2. Sensors & software by CIP (Open source)
IMAGri v2.0 - Multispectral Imaging System
Description:
The Multispectral Acquisiton system
IMAGri v2.0 has a resolution of 640
x 480 pixels.
Remarks:
• Blueprints, software, camera and optics selection are open access.
• Adaptation for different indices (e. g. PRI).
• Can be adapted to any UAV that can carry more than 800 g.
• The upcoming version will incorporate correction for light conditions.

5. NDVI IMAGri v2.0 vs ADC Tetracam
Altitude: 40m
Location: CIP-LIMA

6. ISAM V3.0 – Image stitching software
Description:
- Stitches 2 or more images into one (mosaic)
- Tests were performed with TETRACAM Micro, Snap, ADC (3 bands) and IMAGri (2 bands).
- Current Version can join 5 bands MCA TETRACAM images
Remarks: Source code, final product, tutorials and sample image are free access in our website - GNU GPL License

7. 3. Examples of applications at CIP
(1) Evaluating potato genotypes´ tolerance to
drought

8. (2) Estimating yield
36 days after planting (d.a.p)
45 d.a.p
84 d.a.p.

9. Spatial yield monitoring and prediction
NDVI

10. (3) Pests and diseases
Wavelenght (nm)
300 400 500 600 700 800 900
Reflectance(%)
0
10
20
30
40
50
CONTROL. 23dai
PYVVi. 23dai
Reflectance(%) 0
10
20
30
40
CONTROL. 17dai
PYVVi. 17dai
350 400 450 500 550 600 650 700
3
4
5
6
7
8
9
0/50
Greenhouse experiment
with Potato Yellow Vein
Virus
Cassava experiment

11. 4. The Community of Practice - CoP
Why a CoP?
• Identify mechanisms that enable UAV ARSIS as a
sustainable tool;
• Identify actions to achieve the outcome;
• Determine how to implement these actions in
each context?
 What collaborations are necessary?
 What capacities and learning are required, and how
to develop these?
 What opportunities exist to pursue these activities?
 How can constraints be addressed?

12. A diversity of stakeholders in a
Community of Practice for innovation
Innovation Flow & Feedback Loops
• Developers
(Hardware &
Software)
Developing
• Application
Scientists (Exploring
Applications; Field
Experiments)
Exploring
Potential Uses
• Users of the
information
• Enablers
Real world
Applications

13. Key Highlights:
– Costs, accessibility, and user-friendliness
– Involving local institution at different stages is a must
– Stepwise – From simple to complex tools
– Complementarity with satellite imageries
– Multiple crops
– Yield assessment?
– Is it feasible to discriminate varieties?
32 Participants:
– National, Regional and
International institutions
– 5 CGIAR Centers (CIP,
ICRAF, CIAT, IITA and ILRI),
and ICIPE
Inception Workshop – October 2014, Nairobi

14. International Potato Center – UAV–
platform assembled and tested in
East Africa
Communication:
engaged with journalism
to communicate with
the public, and training
videos about UAV-ARSIS.
Workshop II
June 6-7, 2016
Identifying innovation pathways through participatory processes in
Africa
Stakeholder
group discussions
to identify
outcomes and
actions
Technology
Fair:
Learning
about UAVs
& ARSIS

15. 5. Crop area determination in Tanzania

16. UAV and accessories
1Check the propellers and drone's motors
2Drone's receiver
3Drone's MicroSD
4Frame
5Tester for batteries
6Conector "Y" for batteries
7Fully charged battery of Transmiteer RC
8Fully charged battery of Laptop
9Image from the place
10Waypoints
11Range Extender
12Plastic seats
13Stopwatch
14Sunglasses
Flight with Tetracam
1Camera housing
2TTC camera
3Ground Calibration Target
4Compactflash memory card empty
5Fully charged battery of camera
Flight with MicroTTC
1Camera housing
2 MicroTTC camera
3 Ground Calibration Target
4MicroSD memory card empty
5Fully charged battery (Lipo 850mAh )
Flight with Sony camera
1Camera housing
2Camara Sony
3SD memory card empty
4Fully charged battery of camera
 Site identification and reconnaissance
 Stakeholder participation
 Permits
 Set mission objectives, which
determines:
• Type of UAV
• Sensors
• Image resolutions
• Time to acquire data, e.g. growing
seasons
Pre-mission preparation

17. On-ground preparation and data acquisition
• Assemble and calibrate UAV
• Survey field
• Ground measurement
• Flight plans
• Data acquisition with UAV

18. Data acquisition
Field data processing

19. Land use/cover in Kilosa, Tanzania (sample=100 ha)
Class Area m2
Banana 2246.37
Bare soil 22418.64
Bush/Shrub 165016.74
Cowpeas 10282.3
Fallow 18552.65
Flooded land 19001.4
Paddy Rice 347.93
Grass 46334.72
Homestead 7001.58
Maize&Sesame 1194.4
Maize 85613.98
Maize&Beans 1664.7
Maize&PigeonPeas 21110.6
Maize&SunFlower 13434.66
No data 16958.61
Pigeon Pea 37540.3
Rice 1270.89
Road 25954.66
Sesame 401975.84
Sesame&Pigeon 2311.95
Sunflower 189.69
Recently planted land 51084.71
Water 2406.1

20. 6. Key messages
• UAV-based technologies have a huge potential
in agricultural sector
• Quality of products generated is strongly
dependent on the quality of sensors and data
processing techniques used.
• Establishing a CoP is an excellent approach for
developing technologies that respond to local
needs for enhanced adoption and sustainability
• These technologies will only work in the region if
there is an enabling environment (policies and
capacity building) – REGIONAL HUB

21. RPAS-based remote sensing supporting
agriculture
CIP provides tailor-made solutions for the utilization
of Remotely Piloted Aircraft Systems suited with
required sensors and open source software to
register images, process data and generate the
information needed for phenotyping and timely
agricultural decision making.
Research Team:
Roberto Quiroz
Adolfo Posadas
Hildo Loayza
Corinne Valdivia
Susan Palacios
Mario Balcázar
Luis Silva
Mariella Carbajal
Percy Zorogastúa
Felipe de Mendiburu
Elijah Cheruiyot
Arnold Bett
Dieudonné Harahagazwe
Rodrigo Morales
Mariana Cruz
Carolina Barreda
Software:
http://cipotato.org/csicc/download
Videos:
http://cipotato.org/csicc/videos
Thanks for your attention
www.uav4ag.org
Contact: Roberto Quiroz at
r.quiroz@cgiar.org