Premium
Broad‐Scale Monitoring of Tillage Practices Using Sequential Landsat Imagery
Author(s) -
Zheng Baojuan,
Campbell James B.,
Shao Yang,
Wynne Randolph H.
Publication year - 2013
Publication title -
soil science society of america journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.836
H-Index - 168
eISSN - 1435-0661
pISSN - 0361-5995
DOI - 10.2136/sssaj2013.03.0108
Subject(s) - scale (ratio) , tillage , remote sensing , environmental science , geology , geography , cartography , agronomy , biology
Crop residue management is an important conservation practice that can preserve soil and water resources. Monitoring how farmers manage their land in terms of tillage practices can provide insights on the effectiveness of the conservation programs so that we can improve the implementation of conservation tillage practices to optimize crop yields and environmental conservation. Remote sensing is a cost‐effective and time‐efficient tool for tillage monitoring. Previous studies have established the minimum Normalized Difference Tillage Index (minNDTI) technique at a pixel level for tillage mapping by extracting the minimum values from time‐series Landsat NDTI spectral profiles. The objective of this study was to evaluate the minNDTI technique and to devise strategies for mapping tillage practices at the field level. We incorporated a multiscale image segmentation approach to fill Landsat 7 scan line corrector (SLC)‐off data gaps to facilitate site‐specific and broad‐scale mapping of tillage practices. The gap‐filled Landsat 7 images were combined with Landsat 5 imagery to develop a multitemporal image series for tillage mapping. We then applied object‐based approaches and utilized the USDA cropland data layer to produce field‐level tillage maps, which were evaluated using field observations and Conservation Technology Information Center (CTIC) county‐level tillage data. Our results show that the gap‐filling procedure is effective for the purpose of tillage mapping. Overall classification accuracies of tillage maps range from 69 to 79%. Comparison between remotely sensed and CTIC tillage data suggests that we can monitor tillage practices systematically using remote sensing imagery.