Data mining techniques for identification of spectrally homogeneous areas using NDVI temporal profiles of soybean crop

Abstract

The aim of this study was to group temporal profiles of 10-day composites NDVI product by similarity, which was obtained by the SPOT Vegetation sensor, for municipalities with high soybean production in the state of Paraná, Brazil, in the 2005/2006 cropping season. Data mining is a valuable tool that allows extracting knowledge from a database, identifying valid, new, potentially useful and understandable patterns. Therefore, it was used the methods for clusters generation by means of the algorithms K-Means, MAXVER and DBSCAN, implemented in the WEKA software package. Clusters were created based on the average temporal profiles of NDVI of the 277 municipalities with high soybean production in the state and the best results were found with the K-Means algorithm, grouping the municipalities into six clusters, considering the period from the beginning of October until the end of March, which is equivalent to the crop vegetative cycle. Half of the generated clusters presented spectro-temporal pattern, a characteristic of soybeans and were mostly under the soybean belt in the state of Paraná, which shows good results that were obtained with the proposed methodology as for identification of homogeneous areas. These results will be useful for the creation of regional soybean "masks" to estimate the planted area for this crop.

Crop Data Retrieval Using Earth Observation Data to Support Agricultural Water Management

“…The analysis of just one or a few parcels for each type of crop might not allow for the identification of the standard pattern for each crop along its growth cycle, for one given region, because of noise and clouds in the satellite images and the lack of representativity. Thus, it is more useful to consider a broad set of crop parcels, with similar phenological behaviour, to extract average crop data values, as performed by Esquerdo et al (2011) and Johann et al (2013Johann et al ( , 2016, which are more representative and reliable than those individually obtained, reducing uncertainty resulting from sensor errors and cloud cover and improving data reliability when IWR is calculated at a regional or irrigation perimeter level. Hence, two different approaches were adopted to retrieve crop data as follows:…”

“…Crop data can be estimated using acquired earth observation (EO) data, along the crop growth cycle, at time intervals suitable for the detection of changes in crop phenology (D'Urso & Calera Belmonte, 2006;Vilar, 2015;Vilar et al, 2015;Navarro et al, 2016;Rolim et al, 2016). Presently, the availability of free and open access to high spatial resolution EO data with a short revisit time allows for accurate crop parameter estimation as well as crop growth cycle characterization, improving the identification of each growth cycle stage, which is often imperceptible when lower temporal resolution data are used (El Hajj et al, 2009;D'Urso et al, 2010;Ramme et al, 2010;Johann et al, 2013;Johann et al, 2016;Navarro et al, 2016;Rolim et al, 2016;Grzegozewski et al, 2017;Toureiro et al, 2017). EO methodologies have been widely used for crop evapotranspiration (ETc) and IWR estimation because of the reflective properties of vegetation that allow one to estimate crop biophysical parameters and plant processes such as transpiration (Neale et al, 1989;Calera Belmonte et al, 2005;D'Urso et al, 2010;Paço et al, 2014;Vuolo et al, 2015;Ferreira et al, 2016;Oliveira et al, 2016).…”

Uso de imagens do sensor orbital modis na estimação de datas do ciclo de desenvolvimento da cultura da soja para o estado do Paraná – Brasil

“…Nesta pesquisa, usaram-se também dados de EVI (HUETE et al, 2002), provenientes do produto MOD13Q1 do tile h13v11, com 250 metros de resolução espacial (NASA, 2014), da série temporal, entre os dias 225 (ano de 2011) e 113 (ano de 2012) do calendário juliano, ou seja, de 13-08-2011 a 22-04-2012 do calendário gregoriano, contabilizando 17 imagens para contemplar todo o ciclo de desenvolvimento da cultura da soja (da pré-semeadura até à colheita) no Estado. Foram gerados os perfis espectrotemporais médios de EVI para os 351 municípios, como exemplifica a Figura 2 para o município de Guaíra (Figura 1), considerando apenas os pixels do mapeamento da soja dentro do perímetro de cada município (Figura 1), seguindo a metodologia adotada por JOHANN et al (2013). Este procedimento foi operacionalizado com um sistema de extração de dados de imagens em linguagem de programação "interactive data language" (IDL), desenvolvido por ESQUERDO et al (2011).…”

Data Mining for the Assessment of Management Areas in Precision Agriculture

“…If compared to the algorithm K-means, EM algorithm results had greater differences for MAs with 3 (Figures 2b and 3b) and 4 (Figures 2c and 3c) clusters. This fact is which is justified by the heuristics used by each grouping algorithm (K-means and EM) to generate the clusters, which had already been observed by Johann et al (2013). Although Weka does not offer the error rate of the groupings conducted with EM, it enables visualizing the distribution of the points among the clusters (Table 3).…”