improved potato monitoring in belgium using …...introduction potato processors, traders and...

VITO NV | Boeretang 200 | 2400 MOL | Belgium | Tel. + 32 14 33 55 11 | www.vito.be

LAND USE

Fig.3. Yield differences between participating companies. Circle is variety Nicola , a potato for the fresh market.

Improved potato monitoring in

Belgium using remote sensing

and crop growth modelling

I. Piccard 1, A. Gobin 1, Y. Curnel 2, J.-P. Goffart 2 , V. Planchon 2 , J. Wellens³, B.

Tychon3, N. Cattoor4, R. Cools 4

1 Vlaamse Instelling voor Technologisch Onderzoek (VITO), Mol, BELGIUM 2Centre wallon de Recherches agronomiques (CRA-W), Gembloux, BELGIUM 3 Université de Liège (Arlon), BELGIUM 4 Belgapom, Berlare, BELGIUM

INTRODUCTION

Potato processors, traders and packers largely work with potato contracts. The close follow up of contracted parcels is important to improve the quantity and quality of the crop and reduce risks related to storage, packaging or processing. The use of geo-information by the sector is limited, notwithstanding the great benefits that this type of information may offer. At the same time, new sensor-based equipment continues to gain importance and farmers increasingly invest in this new technology.

The iPot project, financed by the Belgian Science Policy Office (Belspo), aims at providing the Belgian potato processing sector, represented by Belgapom, with near real time information on field condition (weather-soil), crop development and yield estimates, derived from a combination of satellite images and crop growth models.

REMOTE SENSING

A multi-scale approach is used for potato monitoring and yield estimation, integrating field observations and close range sensing measurements with UAV and satellite images and crop growth models.

During the cropping season regular UAV flights (RGB, 3x3 cm) and high resolution satellite images (DMC/Deimos, 22m pixel size) were combined to elucidate crop performance at variety trials. UAV images were processed using a K-means clustering algorithm to classify the crop according to its greenness at 5m resolution. Vegetation indices such as %Cover and LAI were calculated with the Cyclopes algorithm (INRA-EMMAH) on the DMC images (Fig.1).

Fiel

d o

bse

rvat

ion

s U

AV

(2

cm)

Dei

mo

s-1

sat

ellit

e (2

5m

)

% cover

16 May 27 May 6 June 17 June

5m

375m (6 June) (18 May) (23 June) (31 May)

UAV classification result: (green = potatoes)

UAV @ 25m resolution:

Deimos-1 satellite @ 25m resolution:

RESULTS

Both DMC and UAV-based cover maps showed similar patterns, and helped detect different crop stages during the season (Fig.2).

CONCLUSION

The combination of geo-information and crop modelling strengthens the competitiveness of the Belgian potato sector in a global market. A web-based geo-information platform with different dashboards is developed to allow the Belgian potato industry and research centres to access, analyse and combine the data with their own field observations for improved decision-making.

%cover from UAV @ 5m resolution

PHENOLOGY

Understanding and predicting crop phenology and canopy development is important for timely crop management and ultimately for yield estimates. A wide spread field monitoring campaign with crop observations and measurements allowed for further calibration of the satellite image derived vegetation indices. Curve fitting techniques and phenological models were developed and compared with the vegetation indices during the season, both at the trials and farmers’ fields.

Funded by BELSPO

FIELD CAMPAIGN:

- Location of 75 farmers’ fields

- 10 participating potato processing companies

- Varieties: Nicola, Fontane and Bintje

Sensefly eBee (fixed wing)

Daily growth of potatoes between 6 and 17 June 2014

9x: Senescence

8x: Ripening of fruit and seed

7x: Development of fruit

6x: Flowering

5x: Inflorescence emergence

4x: Tuber formation

3x: Main stem elongation (crop cover)

2x: Formation of main stem

1x: Leaf development

0x: Sprouting/Germination

fAPA

R f

rom

Deim

os

Phenolo

gic

al Sta

ge

Days after planting Days after planting

#tubers

Fig.5. iPOT webtool, composed of different dashboards to enable field location and extraction of satellite information.

Fig.6. Phenological development of Bintje and Fontane (top) and correspondence of field phenology observations with satellite derived fAPAR (bottom).

Fig. 2. DMC and UAV-based cover maps correspond to field observations during the growing season.

Fig.1. UAV-based cover maps demonstrate growth of cover during the growing season.