REFINEMENT OF CROPLAND DATA LAYER USING MACHINE LEARNING

Abstract

As the most widely used crop-specific land use data, the Cropland Data Layer (CDL) product covers the entire Contiguous United States (CONUS) at 30-meter spatial resolution with very high accuracy up to 95% for major crop types (i.e., Corn, Soybean) in major crop area. However, the quality of early-year CDL products were not as good as the recent ones. There are many erroneous pixels in the early-year CDL product due to the cloud cover of the original Landsat images, which affect many follow-on researches and applications. To address this issue, we explore the feasibility of using machine learning technology to refine and correct misclassified pixels in the historical CDLs in this study. An end-to-end deep learning-based framework for restoration of misclassified pixels in CDL image is developed and tested. By feeding the CDL time series into the artificial neural network, a crop sequence model is trained and the misclassified pixels in an original CDL map can be restored. In the experiment with the 2005 CDL data of the State of Illinois, the misclassified pixels over Agricultural Statistics Districts (ASD) #1760 were corrected with a reasonable accuracy (> 85%). The findings suggest that the proposed method provides a low-cost and reliable way to refine the historical CDL data, which can be potentially scaled up to the entire CONUS.

Full text

REFINEMENT OF CROPLAND DATA LAYER USING MACHINE LEARNING
Chen Zhang 1,2, Zhengwei Yang 3, Liping Di 1,2,∗
, Li Lin 1,2, Pengyu Hao 1
1Center for Spatial Science and Systems, George Mason University, Fairfax, VA 22030, USA -
(czhang11, ldi, llin2, phao)@gmu.edu
2Department of Geography and Geoinformation Science, George Mason University, Fairfax, VA 22030, USA
3Research and Development Division, U.S. Department of Agriculture National Agricultural Statistics Service,
Washington, DC 20250, USA - Zhengwei.Yang@usda.gov
KEY WORDS: Cropland Data Layer, Machine Learning, Misclassification Correction, Crop Sequence Modeling, Raster Map
Refinement
ABSTRACT:
As the most widely used crop-specific land use data, the Cropland Data Layer (CDL) product covers the entire Contiguous United
States (CONUS) at 30-meter spatial resolution with very high accuracy up to 95% for major crop types (i.e., Corn, Soybean)
in major crop area. However, the quality of early-year CDL products were not as good as the recent ones. There are many
erroneous pixels in the early-year CDL product due to the cloud cover of the original Landsat images, which affect many follow-on
researches and applications. To address this issue, we explore the feasibility of using machine learning technology to refine and
correct misclassified pixels in the historical CDLs in this study. An end-to-end deep learning-based framework for restoration of
misclassified pixels in CDL image is developed and tested. By feeding the CDL time series into the artificial neural network, a crop
sequence model is trained and the misclassified pixels in an original CDL map can be restored. In the experiment with the 2005
CDL data of the State of Illinois, the misclassified pixels over Agricultural Statistics Districts (ASD) #1760 were corrected with
a reasonable accuracy (>85%). The findings suggest that the proposed method provides a low-cost and reliable way to refine the
historical CDL data, which can be potentially scaled up to the entire CONUS.
1. INTRODUCTION
Since its first release of a full state wide data product in 1997,
the Cropland Data Layer (CDL) product of the U.S. Depart-
ment of Agriculture (USDA) National Agricultural Statistics
Service (NASS) has been widely used by growers, agricultural
industry, governments, educators and students, and researchers
world-wide for crop production, agricultural production plan-
ning and management, government policy formulation and de-
cision making, teaching, and various research activities (Liknes
et al., 2009; Thompson, Prokopy; Hao et al., 2015; Lark et al.,
2015; Di et al., 2017). Currently, the CDL data covers the entire
conterminous United States (CONUS) at 30-meter spatial resol-
ution with a high accuracy up to 95% for classifying major crop
types (i.e., Corn, Soybean, and Wheat). However, the quality of
the early-year CDL products was not as good as recent years.
In early years, there are many misclassified pixels in the CDL
products because of cloud cover and lack of satellite images.
Moreover, only a few states of CDL data were produced before
2008. For example, the year 2000 CDL covers only Illinois,
Indiana, Mississippi, North Dakota, and a part of Arkansas and
Iowa. Obviously, the earlier year CDLs’ availability and low
quality issues affect many follow-on Land Use and Land Cover
(LULC) related researches and applications. Therefore, an ef-
fective method for refining and correcting the old CDL data is
badly needed to improve the quality and accuracy of the histor-
ical CDL data.
It is well known that monocropping will result in degradation of
soil, build-up of diseases and pests, and decline in productivity.
Thus crop rotation becomes a common farming practice in U.S.
Corn Belt. The crop rotation can significantly improve the soil
∗Corresponding author
condition, such as fertility and soil physical/chemical proper-
ties (Pikul et al., 2001; Karlen et al., 2006; Govaerts et al., 2007;
Karlen et al., 2013; Van Eerd et al., 2014). Meanwhile, the
crop sequence and cropping decision also have significant im-
pact on crop yields and profitability (Temperly, Borges; Parajuli
et al., 2013; Farmaha et al., 2016). Based on this common crop-
ping practice, many crop mapping and yield estimation mod-
els and approaches were developed. Secchi et al. (2011) con-
structed an prediction model of future land use scenario in the
state of Iowa based on the corn-soybean rotation and production
costs. Sch¨
onhart et al. (2011) developed a crop sequence model
to generate crop rotations based on agronomic criteria and ob-
served data. Sahajpal et al. (2014) detected the pronounced shifts
from grassland to cultivated area by modelling crop rotation in
the U.S. Western Corn Belt. Hao et al. (2016) explored the crop
classification based on the previous-year crop knowledge. Zhang
et al. (2019a) produced a crop cover map of Nebraska State
based on the common crop rotation patterns of corn, soybeans,
winter wheat, and alfalfa. They further implemented a crop
sequence-based machine learning framework for prediction of
crop cover maps (Zhang et al., 2019b).
In this paper, we present a machine learning-based crop se-
quence model to refine the historical CDL data. The proposed
model utilizes artificial neural network (ANN) to automatically
learn crop sequence information from the CDL time series. The
misclassified pixels in the crop cover map can be automatically
identified and corrected using the trained model on the histor-
ical CDL.
The rest of the paper is organized as follows. Section 2 in-
troduces the CDL data, the study area, and an end-to-end ma-
chine learning framework for the historical CDL data refine-
ment. Section 3 demonstrates the experiment results and as-
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3/W11, 2020
PECORA 21/ISRSE 38 Joint Meeting, 6–11 October 2019, Baltimore, Maryland, USA
This contribution has been peer-reviewed.
https://doi.org/10.5194/isprs-archives-XLII-3-W11-161-2020 | © Authors 2020. CC BY 4.0 License.
161

sesses the refinement performance. Section 4 discusses the lim-
itation of the current implementation and gives the conclusion.
2. METHODS
2.1 Cropland Data Layer
CDL is a raster formatted, geo-referenced, crop-specific land
cover map produced by USDA NASS. It is an annual product
covering the entire CONUS at 30-meter spatial resolution from
2008 to present and some states from 1997 to 2007. The pro-
duction of CDL is mainly based on moderate resolution satellite
imagery and extensive agricultural ground truth (Boryan et al.,
2011). The misclassified pixels in the CDL refer to the pixels
that are covered with “clouds” or “no data”. These pixels are
mainly existing in the CDL products before 2006 due to lack
of high-quality satellite data and the algorithm limitation back
then. Examples of the misclassified pixels in the early-year
CDL are shown in Figure 1.
The CDL data products are freely downloaded from CropScape
(https://nassgeodata.gmu.edu/CropScape/), which is developed
and maintained in cooperation with Center for Spatial Informa-
tion Science and Systems of George Mason University (Han et
al., 2012; Zhang et al., 2019c). It provides an easy-to-use Web
GIS application to visualize, analyse, and download CDL data.
All data hosted on CropScape are disseminated via the OGC
standards-compliant geospatial Web services, such as Web Map
Service (WMS), Web Coverage Service (WCS), Web Feature
Service (WFS), and Web Processing Service (WPS).
2.2 Study Area
The Agricultural Statistics District (ASD) #1760 of Illinois state
is selected as the study area. The study area lies on the Central
Corn Belt Plains Ecoregion, which is mainly covered by corn,
soybeans, grassland, and forest as shown in Figure 2. It can
be seen that the 2005 CDL contains a considerable number of
pixels are labelled as “clouds or no data” over the study area.
The purpose of this study is to restore those misclassified pixels
in the study area of 2005 CDL using the machine-learned crop
sequence model.
2.3 Machine Learning Framework
To automatically correct the misclassified pixels in CDL, an
end-to-end machine learning framework is proposed in this pa-
per. The proposed framework is composed of four major com-
ponents: data preparation, model training, classification, and
evaluation.
2.3.1 Data Preparation: In data preparation, the CDLs from
2006 to 2018 are stacked sequentially to form CDL time series.
All pixels of the CDL time series are arranged into a 2-D ar-
ray of samples. Each row of the data set array represents a
pixel consisting of a sequence of crop type values of different
years. Training and validation data sets are randomly sampled
from the “good pixels” in the study area and labelled with 2005
CDL. The experiment data set includes all pixels corresponding
to those misclassified pixels in the study area without labels.
Figure 1. Examples of misclassified pixels in the early-year
CDL data.
Figure 2. Study area with 2005 CDL as the experiment data
(data available from CropScape).
2.3.2 Model Training: The crop sequence model is trained
by feeding the training set into the artificial neural network,
which contains one input layer, multiple hidden layers, and one
output layer. The input layer contains a group of neurons cor-
responding to the same pixel of the CDL time series. Each input
pixel represents a specific value of its crop type. There are mul-
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3/W11, 2020
PECORA 21/ISRSE 38 Joint Meeting, 6–11 October 2019, Baltimore, Maryland, USA
This contribution has been peer-reviewed.
https://doi.org/10.5194/isprs-archives-XLII-3-W11-161-2020 | © Authors 2020. CC BY 4.0 License.
162

Figure 3. Applications of machine learning-based crop sequence model.
tiple hidden layers between the input layer and the output layer.
The output layer uses SoftMax to estimate the probability of
each crop type.
2.3.3 Classification and Validation: By feeding the exper-
iment data set to the well-trained crop sequence model, the mis-
classified pixels in the original CDL can be refined. To validate
the refinement performance, we applied the same crop sequence
model to the validation set. Then we measured the model by
calculating the agreement of the classified label and the original
label of the validation set.
The applications of the proposed machine learning-based crop
sequence model are illustrated in Figure 3. In this study, the
crop sequence model is used to restore the historical crop cover
map. This model, on the other hand, can be also applied to pre-
dict the future crop cover maps with the high-confident training
samples for early-season and in-season crop mapping.
3. RESULTS
The refined 2005 CDL data of ASD #1760 is illustrated in Fig-
ure 4. Comparing the refined result with the original 2005 CDL
data, we observed that the misclassified pixels had been correc-
ted with the crop sequence information learned from the histor-
ical CDL time series.
Figure 4. Refined 2005 CDL of ASD #1760.
The overall accuracy of the refined pixels is unable to be ac-
cessed directly due to lack of ground reference data. Instead, we
utilized the validation data set, derived from the “good pixels”
in the study area of 2005 CDL to indirectly measure the per-
formance of the model. The overall accuracy of validation based
on the validation sample set is over 85%. Therefore, the actual
overall accuracy of the refined pixels may vary. To further val-
idate the performance of refinement, the ground reference data
are required.
4. CONCLUSION
This study investigated the feasibility of using machine learn-
ing technology to refine CDL data. An end-to-end ANN-based
framework was proposed and tested to correct the misclassified
pixels in the historical CDL data. The preliminary experiment
result indicates that the misclassified pixels over the ASD #1760
could be corrected with reasonable accuracy (>85%). The find-
ings suggest that the proposed machine learning approach is ef-
fective and low-cost for correcting the misclassified pixels, and
has great potential for refining the historical CDL over large
geographic area. More experiments and validation will be con-
ducted in the future.
ACKNOWLEDGEMENTS
This work is supported by the U.S. Department of Agriculture
National Agricultural Statistics Service.
REFERENCES
Boryan, C., Yang, Z., Mueller, R., Craig, M., 2011. Mon-
itoring US Agriculture: The US Department of Agricul-
ture, National Agricultural Statistics Service, Cropland Data
Layer Program. Geocarto International, 26(5), 341–358.
doi:10.1080/10106049.2011.562309.
Di, L., Yu, E. G., Kang, L., Shrestha, R., Bai, Y., 2017. RF-
CLASS: A Remote-Sensing-Based Flood Crop Loss Assess-
ment Cyber-Service System for Supporting Crop Statistics
and Insurance Decision-Making. Journal of Integrative Agri-
culture, 16(2), 408-423. doi:10.1016/S2095-3119(16)61499-
5.
Farmaha, B. S., Eskridge, K. M., Cassman, K. G., Specht,
J. E., Yang, H., Grassini, P., 2016. Rotation Impact on On-
Farm Yield and Input-Use Efficiency in High-Yield Irrigated
Maize–Soybean Systems. Agronomy Journal, 108(6), 2313-
2321. doi:10.2134/agronj2016.01.0046.
Govaerts, B., Mezzalama, M., Unno, Y., Sayre, K. D.,
Luna-Guido, M., Vanherck, K., Dendooven, L., Deck-
ers, J., 2007. Influence of Tillage, Residue Manage-
ment, and Crop Rotation on Soil Microbial Biomass and
Catabolic Diversity. Applied Soil Ecology, 37(1), 18-30.
doi:10.1016/j.apsoil.2007.03.006.
Han, W., Yang, Z., Di, L., Mueller, R., 2012. CropScape: A
Web Service Based Application for Exploring and Dissemin-
ating US Conterminous Geospatial Cropland Data Products
for Decision Support. Computers and Electronics in Agricul-
ture, 84, 111–123. doi:10.1016/j.compag.2012.03.005.
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3/W11, 2020
PECORA 21/ISRSE 38 Joint Meeting, 6–11 October 2019, Baltimore, Maryland, USA
This contribution has been peer-reviewed.
https://doi.org/10.5194/isprs-archives-XLII-3-W11-161-2020 | © Authors 2020. CC BY 4.0 License.
163

Hao, P., Wang, L., Zhan, Y., Wang, C., Niu, Z., Wu, M.,
2016. Crop Classification Using Crop Knowledge of the
Previous-Year: Case Study in Southwest Kansas, USA.
European Journal of Remote Sensing, 49(1), 1061-1077.
doi:10.5721/EuJRS20164954.
Hao, P., Zhan, Y., Wang, L., Niu, Z., Shakir, M.,
2015. Feature Selection of Time Series MODIS Data for
Early Crop Classification Using Random Forest: A Case
Study in Kansas, USA. Remote Sensing, 7(5), 5347-5369.
doi:10.3390/rs70505347.
Karlen, D. L., Cambardella, C. A., Kovar, J. L., Colvin, T. S.,
2013. Soil Quality Response to Long-Term Tillage and Crop
Rotation Practices. Soil and Tillage Research, 133, 54-64.
doi:10.1016/j.still.2013.05.013.
Karlen, D. L., Hurley, E. G., Andrews, S. S., Cambardella,
C. A., Meek, D. W., Duffy, M. D., Mallarino, A. P., 2006.
Crop Rotation Effects on Soil Quality at Three Northern
Corn/Soybean Belt Locations. Agronomy Journal, 98(3),
484-495. doi:10.2134/agronj2005.0098.
Lark, T. J., Salmon, J. M., Gibbs, H. K., 2015. Cropland
Expansion Outpaces Agricultural and Biofuel Policies in
the United States. Environmental Research Letters, 10(4),
044003. doi:10.1088/1748-9326/10/4/044003.
Liknes, G. C., Nelson, M. D., Gormanson, D. D., Hansen,
M., 2009. The Utility of the Cropland Data Layer for
Forest Inventory and Analysis. Proceedings of the Eighth
Annual Forest Inventory and Analysis Symposium, 259–264.
doi:10.2737/WO-GTR-79.
Parajuli, P. B., Jayakody, P., Sassenrath, G. F., Ouyang, Y., Pote,
J. W., 2013. Assessing the Impacts of Crop-Rotation and
Tillage on Crop Yields and Sediment Yield Using a Model-
ing Approach. Agricultural Water Management, 119, 32-42.
doi:10.1016/j.agwat.2012.12.010.
Pikul, J. L., Carpenter-Boggs, L., Vigil, M., Schumacher, T. E.,
Lindstrom, M. J., Riedell, W. E., 2001. Crop Yield and Soil
Condition under Ridge and Chisel-Plow Tillage in the North-
ern Corn Belt, USA. Soil and Tillage Research, 60(1), 21-33.
doi:10.1016/S0167-1987(01)00174-X.
Sahajpal, R., Zhang, X., Izaurralde, R. C., Gelfand, I., Hurtt,
G. C., 2014. Identifying Representative Crop Rotation Pat-
terns and Grassland Loss in the US Western Corn Belt.
Computers and Electronics in Agriculture, 108, 173-182.
doi:10.1016/j.compag.2014.08.005.
Sch¨
onhart, M., Schmid, E., Schneider, U. A., 2011. CropRota –
A Crop Rotation Model to Support Integrated Land Use As-
sessments. European Journal of Agronomy, 34(4), 263-277.
doi:10.1016/j.eja.2011.02.004.
Secchi, S., Kurkalova, L., Gassman, P. W., Hart, C., 2011.
Land Use Change in a Biofuels Hotspot: The Case of
Iowa, USA. Biomass and Bioenergy, 35(6), 2391-2400.
doi:10.1016/j.biombioe.2010.08.047.
Temperly, R. J., Borges, R., 2006. Tillage and Crop Rotation
Impact on Soybean Grain Yield and Composition. Agronomy
Journal, 98(4), 999-1004. doi:10.2134/agronj2005.0215.
Thompson, A. W., Prokopy, L. S., 2009. Tracking Urban
Sprawl: Using Spatial Data to Inform Farmland Pre-
servation Policy. Land Use Policy, 26(2), 194-202.
doi:10.1016/j.landusepol.2008.02.005.
Van Eerd, L. L., Congreves, K. A., Hayes, A., Verhallen, A.,
Hooker, D. C., 2014. Long-Term Tillage and Crop Rota-
tion Effects on Soil Quality, Organic Carbon, and Total Ni-
trogen. Canadian Journal of Soil Science, 94(3), 303-315.
doi:10.4141/cjss2013-093.
Zhang, C., Di, L., Lin, L., Guo, L., 2019a. Ex-
tracting Trusted Pixels from Historical Cropland Data
Layer Using Crop Rotation Patterns: A Case Study
in Nebraska, USA. 2019 8th International Confer-
ence on Agro-Geoinformatics (Agro-Geoinformatics), 1–6.
doi:10.1109/Agro-Geoinformatics.2019.8820236.
Zhang, C., Di, L., Lin, L., Guo, L., 2019b. Machine-
Learned Prediction of Annual Crop Planting in the U.S.
Corn Belt Based on Historical Crop Planting Maps.
Computers and Electronics in Agriculture, 166, 104989.
doi:10.1016/j.compag.2019.104989.
Zhang, C., Di, L., Yang, Z., Lin, L., Yu, E. G., Yu, Z., Rahman,
M. S., Zhao, H., 2019c. Cloud Environment for Dissemin-
ating NASS Cropland Data Layer. 2019 8th International
Conference on Agro-Geoinformatics (Agro-Geoinformatics),
1–5. doi:10.1109/Agro-Geoinformatics.2019.8820465.
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3/W11, 2020
PECORA 21/ISRSE 38 Joint Meeting, 6–11 October 2019, Baltimore, Maryland, USA
This contribution has been peer-reviewed.
https://doi.org/10.5194/isprs-archives-XLII-3-W11-161-2020 | © Authors 2020. CC BY 4.0 License.
164