XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
Table 2. Accuracy assessment of summer 2010 classification
(major classes only)
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B7, 2012
crop types. For summer, cotton and sorghum were considered
as the major crops and crops like sunflower, mung beans,
of 78% (k = 0.7) (Table 4).
millets and fodder crops were grouped into a class called other
crops. The SVM model generated classified the image into
these broadacre crop types (Figure 6) with an overall accuracy
Classes Reference
Fallow Crop Pasture Woody Total
Fallow 68 7 1 0 76
= Crop 30 369 23 6 428
3 Pasture 1 0 14 1 16
S Woody 0 0 0 38 38
© Total 99 376 38 45 558
Overall accuracy = 87%
Kappa coefficient (k) = 0.73
Lower classification accuracy in the case of summer data could
be attributed mainly to two reasons. During summer, there was
more classification error between crop and pasture. High
amount of rainfall during the summer growing season causes a
significant increase in vegetative growth in pasture areas and
this in turn could make it difficult to distinguish these areas
from cropping, spectrally. The second reason could be the
noticeably higher cloud cover during the summer. It may be
noted that EVI range is observed to be the most important input
variable for SVM modelling and cloud-affected pixels could
decrease the number pixels available for EVI range estimation
over the growing season. Further, a separate analysis, carried
out by omitting training data sets over cloud-affected pixels,
indicated the accuracy could be as high as 95 % (k = 0.9). The
(Crop type
HE Cotton
I Sorghum *
RS Other crop:
Figure 6. Classification of broad acre crop types for summer
2010 and winter 2011
Table 4. Accuracy assessment of summer 2010 classification of
MLC was applied on the same dataset and the results clearly crop types
revealed that SVM techniques not only produced superior Classes Reference
classification accuracy, but also generated a neater and i Cotton Fallow Other crops Pasture Sorghum Woody Total
speckle-free image (Figure 5). Cotton 162 6 8 5 16 2 19
EN — Fallow 1 72 4 1 2 0 80
Other crops 0 2 7 0 0 0 9
+ Pasture 0 2 2 20 0 1 25
£ Sorghum 19 17 23 12 132 1 204
& _Woody 0 0 0 0 0 41 41
©. Total 182 99 44 38 150 45 558
Overall accuracy = 78%
Kappa coefficient (k) = 0.70
Similarly, SVM models were generated for separating
broadacre crop types for winter 2011 (Figure 6). Major crop
types identified were barley and wheat. Crops like chick pea,
and fodder were grouped into other crops. Winter crop type
classification accuracy was again found to be slightly higher
SVM - object based classification Maximum Likelihood classification (79%, k= 0.73) than that of summer (Table 3).
Figure 5. Comparison of Support Vector Machines and Table 5. Accuracy assessment of winter 2011 classification of
Maximum Likelihood Classified images crop types
Table 3. Accuracy assessment of winter 2011 classification Class Reference
(major classes only) Bg Barley Fallow Other crops Pasture Wheat Woody _ Total
Barley 14 0 3 0 2 0 20
Classes Reference Fallow 0 70 0 0 0 0 70
Fallow Crop Pasture Woody Total Other crops 6 0 11 0 3 0 20
Fallow 70 0 0 0 70 3 Pasture 3 0 11 178 7 12 211
3 Crop 0 244 4 4 274 = Wheat 30 0 27 6 134 0 197
= Pasture 0 1 158 8 167 3 Woody 0 0 2 0 2 78 80
= Woody 0 1 0 78 79 © Total 50 70 54 184 0 90 598
© "Total 70 246 184 90 590 Overall accuracy = 79%
Overall accuracy = 93%
Kappa coefficient (k) = 0.90
This project aims to develop operational methods for crop type
classification for Queensland. In pursuit of this, a preliminary
investigation was attempted to classify broadacre crop types for
both crop seasons. For summer 2010, the crop class mapped in
the first phase (Table 2 and ) was further classified to different
Kappa coefficient (k) = 0.73
Developing operational methods for the assessment of crop
distribution is crucial for the effective implementation of
agricultural policies. For example, the State Government of
Queensland, Australia, passed the Strategic Cropping Land
(SCL) Bill in December 2011 to protect land that is highly
suitable for cropping, manage the impacts of development on