B8, 2012
0.12. Note that
1 cover of 0.0).
| ue
5 | 070 |
1 07% |
3 | 059 |
2 0.74
green cover on
d Biomass
nding measured
across sites and
rin. kg ha)
1s for dates up
974) plotted by
e the maximum
easured biomass
'ression model:
(l)
. The regression
y NDVI did not
s as it did for
lues. Pooling all
ue of 0.13 and
nsors (ACS 470
ta was subset to
0 Horsham data
comparison was
d the highest
[. One possible
ity of the sensor
ie to the small
en sensor NDVI
biomass. Linear
ysham data for
nclusive), which
: of 115 kg ha”
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B8, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
pe
12000 7
10000 4 = o ACS 470
8000 —
6000 -
4000 +
2000 +
Biomass, as Dry Matter kg ha-1
D
e ACS 210
0.4
Sensor NDVI
Figure 2. Above ground biomass measured as dry matter kg
ha! plotted against the mean NDVI measured by
plot for all data up to the flowering stage, Zadoks 60
(N=1475). The non linear equation fitted yield re
0.27 and a standard error of 2133 kg ha”.
Std. Error (kg ha’!
Sensor N r
All Dates
All Sensors 1858 0.13 3459
ACS 470 1214 0.03 3599
Greenseeker 379 0.30 3017
ACS 210
Until Flowering (Zadoks 60)
All Sensors 1477 0.27 2133
ACS 470 947 0.22 2377
Greenseeker 195 0.37 1537
ACS 210 265 0.64 372
ACS 210 148 0.67 115
Low
biomass*
* Fitted linear regression to data with 0 < DM < 1000 kg ha
Table 3. Regression results for Total Above Ground Dry
Matter (kg ha") on NDVI
4. CONCLUSIONS
The commercially available active optical sensors are robust
and can be used under most sky conditions, so they are of
potential interest for ground truth of biomass for various remote
sensing applications. The results show a strong linear relation
between the sensor NDVI and fractional green cover for a
dataset that included three different sensors and multiple sites
and dates. However, the relation between measured biomass and
sensor NDVI was found to be linear only at low values of
biomass (less than 1000 kg ha), and non-linear regression
resulted in low r? values for the dataset. Results of the study
indicate the sensors may provide good estimates of fractional
green cover, although caution should be applied when directly
comparing NDVI from different sensors models.
5. REFERENCES
Li, Y., Chen, D., Walker, C.N., Angus, J.F., 2010. Estimating
the nitrogen status of crops using a digital camera. Field Crops
Research 118(3), pp. 221-227.
Payne, R.W., Murray, D.A., Harding, S.A., Baird, D.B., Soutar,
D.M., 2009. GenStat for Windows (12th Edition) Introduction.
VSN International, Hemel Hempstead.
Rouse, J.W., Jr., Haas, R. H., Schell, J.A., Deering, D.W, 1973.
Monitoring vegetation systems in the Great Plains with ETRS. ,
Earth Res. Tech. Satellite-1 Symp, Goddard Space Flight Cent.,
Washington, DC, pp. 309-317.
Zadoks, J.C., Chang, T.T., Konzak, C.F., 1974. A decimal code
for the growth stages of cereals. Weed Research 14(6), pp. 415-
421.
6. ACKNOWLEDGEMENTS
The authors wish to thank the Grains Research Development
Corporation for research funding of multiple projects that
supported the data acquisition and analysis of this work. We are
also indebted to our grower collaborators who allowed us to
work on their farms.
