Portal de Periódicos da CAPES

Influence of Gradient Estimation on Clod Identification on a Seedbed Digital Elevation Model

2011; Geological Society of America; Volume: 17; Issue: 4 Linguagem: Inglês

10.2113/gseegeosci.17.4.337

1558-9161

O. Chimi-Chiadjeu, Edwige Vannier, Richard Dusséaux, O. Taconet,

Agronomic Practices and Intercropping Systems

Research Article| November 01, 2011 Influence of Gradient Estimation on Clod Identification on a Seedbed Digital Elevation Model OLIVIER CHIMI-CHIADJEU; OLIVIER CHIMI-CHIADJEU Université de Versailles Saint-Quentin, CNRS/INSU, LATMOS-IPSL, 11 Bd d'Alembert, Guyancourt, 78280 France Search for other works by this author on: GSW Google Scholar EDWIGE VANNIER; EDWIGE VANNIER 1 Université de Versailles Saint-Quentin, CNRS/INSU, LATMOS-IPSL, 11 Bd d'Alembert, Guyancourt, 78280 France 1Corresponding author email: edwige.vannier@latmos.ipsl.fr. Search for other works by this author on: GSW Google Scholar RICHARD DUSSÉAUX; RICHARD DUSSÉAUX Université de Versailles Saint-Quentin, CNRS/INSU, LATMOS-IPSL, 11 Bd d'Alembert, Guyancourt, 78280 France Search for other works by this author on: GSW Google Scholar ODILE TACONET ODILE TACONET Université de Versailles Saint-Quentin, CNRS/INSU, LATMOS-IPSL, 11 Bd d'Alembert, Guyancourt, 78280 France Search for other works by this author on: GSW Google Scholar Author and Article Information OLIVIER CHIMI-CHIADJEU Université de Versailles Saint-Quentin, CNRS/INSU, LATMOS-IPSL, 11 Bd d'Alembert, Guyancourt, 78280 France EDWIGE VANNIER 1 Université de Versailles Saint-Quentin, CNRS/INSU, LATMOS-IPSL, 11 Bd d'Alembert, Guyancourt, 78280 France RICHARD DUSSÉAUX Université de Versailles Saint-Quentin, CNRS/INSU, LATMOS-IPSL, 11 Bd d'Alembert, Guyancourt, 78280 France ODILE TACONET Université de Versailles Saint-Quentin, CNRS/INSU, LATMOS-IPSL, 11 Bd d'Alembert, Guyancourt, 78280 France 1Corresponding author email: edwige.vannier@latmos.ipsl.fr. Publisher: Association of Environmental & Engineering Geologists First Online: 02 Mar 2017 Online ISSN: 1558-9161 Print ISSN: 1078-7275 © 2011 Association of Environmental & Engineering Geologists Environmental & Engineering Geoscience (2011) 17 (4): 337–352. https://doi.org/10.2113/gseegeosci.17.4.337 Article history First Online: 02 Mar 2017 Cite View This Citation Add to Citation Manager Share Icon Share Facebook Twitter LinkedIn MailTo Tools Icon Tools Get Permissions Search Site Citation OLIVIER CHIMI-CHIADJEU, EDWIGE VANNIER, RICHARD DUSSÉAUX, ODILE TACONET; Influence of Gradient Estimation on Clod Identification on a Seedbed Digital Elevation Model. Environmental & Engineering Geoscience 2011;; 17 (4): 337–352. doi: https://doi.org/10.2113/gseegeosci.17.4.337 Download citation file: Ris (Zotero) Refmanager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex toolbar search Search Dropdown Menu toolbar search search input Search input auto suggest filter your search All ContentBy SocietyEnvironmental & Engineering Geoscience Search Advanced Search Abstract The structure of the soil surface is a key parameter to understanding soil properties and physical processes related to water infiltration or runoff and soil erosion. A seedbed surface has a cloddy structure, which is responsible for its random roughness. The present study focuses on seedbed cloddiness characterization. It relies on a new gradient-based method, introduced recently in order to identify automatically the individual clods or large aggregates on a three-dimensional digital elevation model (DEM) of the soil surface. We analyze the sensitivity of clod identification and characterization to the gradient estimation method. For this purpose, the partial derivatives are evaluated by four different methods, and a ground truth of clod boundaries has been made by a soil scientist on two seedbed surfaces (an artificial surface and an in-field one). The gradient method involving an error in the fourth power of the spacing and the greatest number of neighboring pixels in the estimation enables an improvement in the sensitivity of clod identification by up to 10 percent, achieving detection performance greater than 73 percent. Comparing segmented contours with ground truth delineation, good agreement is found for half of the clods regardless of the numerical gradient computation method. For the other half, the area is underestimated because it is not possible to extend the contours. The mean variation of shape parameters is slightly sensitive to the gradient computation, with differences of around 5 percent. When decreasing the DEM resolution from 1 mm to 2 mm, the difference between the four methods of gradient computation is strengthened. You do not have access to this content, please speak to your institutional administrator if you feel you should have access.