Abstract
An accurate estimation of soybean crop areas while the plants are still in the field is highly necessary for reliable calculation of real crop parameters as to yield, production and other data important to decision-making policies related to government planning. An algorithm for soybean classification over the Rio Grande do Sul State, Brazil, was developed as an objective, automated tool. It is based on reflectance from medium spatial resolution images. The classification method was called the RCDA (Reflectance-based Crop Detection Algorithm), which operates through a mathematical combination of multi-temporal optical reflectance data obtained from Landsat-5 TM images. A set of 39 municipalities was analyzed for eight crop years between 1996/1997 and 2009/2010. RCDA estimates were compared to the official estimates of the Brazilian Institute of Geography and Statistics (IBGE) for soybean area at a municipal level. Coefficients R2 were between 0.81 and 0.98, indicating good agreement of the estimates. The RCDA was also compared to a soybean crop map derived from Landsat images for the 2000/2001 crop year, the overall map accuracy was 91.91% and the Kappa Index of Agreement was 0.76. Due to the calculation chain and pre-defined parameters, RCDA is a timesaving procedure and is less subjected to analyst skills for image interpretation. Thus, the RCDA was considered advantageous to provide thematic soybean maps at local and regional scales.
Highlights
Statistical data in agriculture plays a key role in the food supply chain and improvements in methods of precise and timely estimates for crop areas and yield are extremely important for management, trade and pricing policies
Most of the research related to crop area estimation is associated with classification of Landsat Thematic Mapper (TM) images of medium spatial resolution [3]
Soybean area provided by RCDA was estimated by municipality and compared to official estimates provided by IBGE using regression analysis
Summary
Statistical data in agriculture plays a key role in the food supply chain and improvements in methods of precise and timely estimates for crop areas and yield are extremely important for management, trade and pricing policies. The main difficulties were clearly associated with the handling of a large volume of data and a timely generation of the desired products Another two difficulties were the pressure for results and systematic cloud cover occurrences. Focused on the crop area problem in Brazil, several studies [1,9,10] have been conducted using different conceptual approaches with high temporal-resolution data with coarser spatial-resolution These methodologies did not prove to be useful to routine monitoring, and were generally applied for relatively few crop years and specific regions Other developments as made with the MODIS Crop Detection Algorithm (MCDA) [11], produced R2 greater than 0.95 and overall accuracy of 82% for several crop years in Rio
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
