Influência da distância entre passadas de colhedora equipada com monitor de colheita na precisão dos mapas de produtividade na cultura do milho

Abstract

RESUMO:Um dos entraves para a maior expansão da agricultura de precisão está, principalmente, no elevado custo dos equipamentos a serem utilizados. A geração de mapas de produtividade é a fase de mais fácil execução na agricultura de precisão. Entretanto, quando várias colhedoras operam simultaneamente em uma determinada área, nem sempre é possível dispor de monitores de colheita em todas as máquinas. Assim, neste trabalho, foi avaliado o decréscimo da precisão em mapas de produtividade obtidos a partir de áreas onde se utilizaram colhedoras não-equipadas com monitor de colheita associadas a colhedoras monitoradas. Também foram comparados os métodos de interpolação "inverso da distância ao quadrado" e "krigagem". Inicialmente, a interpolação do mapa com todos os pontos de amostragem (original) foi gerada e, a partir dessa, foi feita a eliminação de passadas de modo a obter os mapas simulados. Na comparação entre os mapas simulados e o original, foi utilizado o coeficiente de desvio relativo -CDR, e o índice kappa. O número de passadas eliminadas diminuiu a precisão dos mapas em ambos os métodos de interpolação. Sendo assim, a quantidade mínima necessária de monitores em um conjunto de colhedoras depende da precisão desejada. Portanto, as máquinas devem ser intercaladas de maneira que as amostras sejam coletadas uniformemente por toda a área a ser colhida.PALAVRAS-CHAVE: agricultura de precisão, métodos de interpolação, mapas temáticos. INFLUENCE OF DISTANCE BETWEEN COMBINES EQUIPPED WITH YIELD MONITORS ON THE PRECISION OF YIELD MAPS FOR CORN CROPS ABSTRACT:One of the main obstacles to the larger expansion of precision agriculture is the high cost of the equipments to be used. The generation of yield maps is the easiest step in the sequence of the precision agriculture process. Nevertheless, the availability of yield monitors for all the machines is not always possible when several combines are concomitantly harvesting in a given area. Within this context, this research was carried out aiming to evaluate the decrease of precision of yield maps obtained from data collected in areas where combines not equipped with yield monitors were harvesting concomitantly with combines equipped with such device. Two different interpolation methods, the "inverse of the square distance" and the "kriging", were also evaluated. Initially, the interpolation of the map with all sampling sites (original) was generated, and from that, the elimination of combine steps was performed, therefore originating the simulated maps. To compare the original maps with the simulated ones, the coefficient of relative deviation and the Kappa index were used. The number of eliminated combine steps decreased the precision of yield maps as demonstrated by both interpolation used methods. Therefore, the minimum number of monitors needed in a group of combines depends on the desired precision. Thus, the intercalation of monitored and non-monitored combines should be performed so that the samples are uniformly collected within all the area to be harvested.

Model to estimate the sampling density for establishment of yield mapping

“…For the same sampling density, CRD, deviation from the original map, was smaller for the interpolation inverse of square distance (ISD), followed by the inverse of distance (ID) and lastly by kriging (KRIG) (except for higher density of points). This result is in agreement with Coelho et al (2009) and Bazzi et al (2008), who found better performance for interpolation using inverse distance weighting when compared to kriging. Kappa indexes resulting from comparison between simulated and original grids (Figure 7) decreased progressively as sampling density for the three interpolation methods was reduced.…”

“…As with the CRD, for the same sample density, kappa index showed higher concordance with the original map for the ISD interpolator; followed by ID and finally by Krig (except for the highest density of points). This result is in agreement with and Bazzi et al (2008). Comparing the explanation of the dependent variable kappa index, the R maps performed by both CRD and kappa methods lead to similar results.…”

“…The yield monitor used was AgLeader® PF 3000®, installed in a combine harvester New Holland TC 57®, corn platform with six row. 14,760 points were collected, which passed through filtration and removal of inconsistent data following the methodology used by Bazzi et al (2008). In this analysis, spots with times of filling and emptying the harvester less than six seconds and delay time of less than twelve seconds were eliminated.…”

Comparison of maps of spatial variability of soil resistance to penetration constructed with and without covariables using a spatial linear model

“…The Kappa index (K) (BAZZI et al, 2008, FOODY, 2009FOODY, 2010) has been used to measure the accuracy of thematic classifications, recommended as an appropriate measure of accuracy for using all elements of the error matrix. Provides a measure of agreement between the values of the reference map and the values from the model map, being defined by Equation (4),…”

Redundant variables and the quality of management zones

“…4) The Kappa index has been used to measure the accuracy of thematic classifications (BAZZI et al, 2008;BASTIANI et al, 2012;DALPOSSO et al, 2012) and is recommended as an appropriate measure of accuracy for all elements from the error matrix. It was used to evaluate the spatial agreement among maps of more efficient MZs (Equation 6).…”