Ghent University Academic Bibliography

Advanced

A neural network approach to topsoil clay prediction using an EMI-based soil sensor

Liesbet Cockx UGent, Marc Van Meirvenne UGent, UWA Vitharana, Friedl Vancoillie UGent, Lieven Verbeke UGent, David Simpson UGent and Timothy Saey UGent (2010) Progress in Soil Science. p.245-254
abstract
High-resolution proximal soil sensor data are an important source of information for optimising the prediction of soil properties. On a 10.5 ha arable field, an intensive EM38DD survey with a resolution of 2 m x 2 m resulted in 19,694 measurements of ECa-H and ECa-V. A large textural variation was present in the subsoil due to the presence of former water channels. Nevertheless, both ECa-V and ECa-H data displayed the same spatial variability. This spatial similarity indicated the strong influence of the subsoil heterogeneity on the ECa-H measurements. Using variography, two scales of ECa variability were identified: short-range (similar to 35 m) variability associated with the channel pattern and wider within-field variability (similar to 200 m). Using artificial neural networks (ANNs), prediction of the topsoil clay content was optimised (i) by using an input window size of 3, 5, 7, 9, and 11 pixels to account for local contextual influence and (ii) by including both ECa-H and ECa-V in the network input layer to isolate the response from the topsoil. The models were evaluated using R-2 and the relative mean squared estimation error (rMSEE) of the test data. The most accurate predictions were obtained using both orientations of the EM38DD sensor without contextual information (R-2 = 0.66, rMSEE = 0.40). The importance of ECa-V on the topsoil clay prediction was expressed by a relative improvement of the rMSEE of 29%. For comparison, a multivariate linear regression (MVLR) was performed to predict the topsoil clay content based on the two orientations. The ANN models up to a window size of 5 pixels outperformed the MVLR, which resulted in an R-2 of 0.42 and an rMSEE of 0.63. ANN analysis based on both orientations of the EM38DD appears to be a useful tool to extract topsoil information from depth-integrated EM38DD measurements.
Please use this url to cite or link to this publication:
author
organization
year
type
conference
publication status
published
subject
keyword
Artificial neural networks, YIELD, EM38DD, Topsoil texture, EM38, Prediction, VARIABILITY
in
Progress in Soil Science
editor
Raphael A Viscarra Rossel, Alex B McBratney and Budiman Minasny
issue title
Proximal soil sensing
pages
245 - 254
publisher
Springer
place of publication
Dordrecht, The Netherlands
conference name
1st Global workshop on High Resolution Digital Soil Sensing and Mapping
conference location
Sydney, Australia
conference start
2008-02-05
conference end
2008-02-08
Web of Science type
Proceedings Paper
Web of Science id
000281598600020
ISBN
9789048188598
9789048188581
DOI
10.1007/978-90-481-8859-8_20
language
English
UGent publication?
yes
classification
P1
copyright statement
I have transferred the copyright for this publication to the publisher
id
1027658
handle
http://hdl.handle.net/1854/LU-1027658
date created
2010-08-23 14:28:31
date last changed
2013-04-29 14:15:47
@inproceedings{1027658,
  abstract     = {High-resolution proximal soil sensor data are an important source of information for optimising the prediction of soil properties. On a 10.5 ha arable field, an intensive EM38DD survey with a resolution of 2 m x 2 m resulted in 19,694 measurements of ECa-H and ECa-V. A large textural variation was present in the subsoil due to the presence of former water channels. Nevertheless, both ECa-V and ECa-H data displayed the same spatial variability. This spatial similarity indicated the strong influence of the subsoil heterogeneity on the ECa-H measurements. Using variography, two scales of ECa variability were identified: short-range (similar to 35 m) variability associated with the channel pattern and wider within-field variability (similar to 200 m). Using artificial neural networks (ANNs), prediction of the topsoil clay content was optimised (i) by using an input window size of 3, 5, 7, 9, and 11 pixels to account for local contextual influence and (ii) by including both ECa-H and ECa-V in the network input layer to isolate the response from the topsoil. The models were evaluated using R-2 and the relative mean squared estimation error (rMSEE) of the test data. The most accurate predictions were obtained using both orientations of the EM38DD sensor without contextual information (R-2 = 0.66, rMSEE = 0.40). The importance of ECa-V on the topsoil clay prediction was expressed by a relative improvement of the rMSEE of 29\%. For comparison, a multivariate linear regression (MVLR) was performed to predict the topsoil clay content based on the two orientations. The ANN models up to a window size of 5 pixels outperformed the MVLR, which resulted in an R-2 of 0.42 and an rMSEE of 0.63. ANN analysis based on both orientations of the EM38DD appears to be a useful tool to extract topsoil information from depth-integrated EM38DD measurements.},
  author       = {Cockx, Liesbet and Van Meirvenne, Marc and Vitharana, UWA and Vancoillie, Friedl and Verbeke, Lieven and Simpson, David and Saey, Timothy},
  booktitle    = {Progress in Soil Science},
  editor       = {Viscarra Rossel, Raphael A and McBratney, Alex B and Minasny, Budiman},
  isbn         = {9789048188598},
  keyword      = {Artificial neural networks,YIELD,EM38DD,Topsoil texture,EM38,Prediction,VARIABILITY},
  language     = {eng},
  location     = {Sydney, Australia},
  pages        = {245--254},
  publisher    = {Springer},
  title        = {A neural network approach to topsoil clay prediction using an EMI-based soil sensor},
  url          = {http://dx.doi.org/10.1007/978-90-481-8859-8\_20},
  year         = {2010},
}

Chicago
Cockx, Liesbet, Marc Van Meirvenne, UWA Vitharana, Friedl Vancoillie, Lieven Verbeke, David Simpson, and Timothy Saey. 2010. “A Neural Network Approach to Topsoil Clay Prediction Using an EMI-based Soil Sensor.” In Progress in Soil Science, ed. Raphael A Viscarra Rossel, Alex B McBratney, and Budiman Minasny, 245–254. Dordrecht, The Netherlands: Springer.
APA
Cockx, Liesbet, Van Meirvenne, M., Vitharana, U., Vancoillie, F., Verbeke, L., Simpson, D., & Saey, T. (2010). A neural network approach to topsoil clay prediction using an EMI-based soil sensor. In R. A. Viscarra Rossel, A. B. McBratney, & B. Minasny (Eds.), Progress in Soil Science (pp. 245–254). Presented at the 1st Global workshop on High Resolution Digital Soil Sensing and Mapping, Dordrecht, The Netherlands: Springer.
Vancouver
1.
Cockx L, Van Meirvenne M, Vitharana U, Vancoillie F, Verbeke L, Simpson D, et al. A neural network approach to topsoil clay prediction using an EMI-based soil sensor. In: Viscarra Rossel RA, McBratney AB, Minasny B, editors. Progress in Soil Science. Dordrecht, The Netherlands: Springer; 2010. p. 245–54.
MLA
Cockx, Liesbet, Marc Van Meirvenne, UWA Vitharana, et al. “A Neural Network Approach to Topsoil Clay Prediction Using an EMI-based Soil Sensor.” Progress in Soil Science. Ed. Raphael A Viscarra Rossel, Alex B McBratney, & Budiman Minasny. Dordrecht, The Netherlands: Springer, 2010. 245–254. Print.