Efeitos diretos e indiretos de parâmetros dos semivariogramas referentes aos atributos químicos do solo sobre a produtividade de grãos de soja

In order to add value to soybens crops, and hence the marketing, medium and large producers have been using precision agriculture techniques (PA), as the Remote Sensing, Geographic Information Systems (GIS) and positioning satellite, to assist the management of crops. Thus, given the economic relevance of that culture to the southwest of Paraná State and Brazil, scientific studies to increase their productivity and profitability are of main importance. The objective of this study was to evaluate the correlation between the chemical soil properties and soybean yield for each estimated parameter of semivariogram (range, nugget and level effect), and the deployment of these correlations in direct and indirect effects, aiming to improve the mapping process of spatial variability of soil chemical properties for use in PA. The hypothesis is that not all attributes of soil used to estimate the semivariogram parameters has a direct effect on productivity, and that even in groups of plants within a larger area it is possible to estimate the parameters of the semivariograms. The experiment was conducted in a commercial area of 19.7 ha, located in the city of Pato Branco - PR, central geographic coordinates 26º 11 '35 "South, 52 43' 05" West longitude, and average altitude of 780 m. The area is planted with soybeans for over 30 years, currently being adopted to cultivate Brasmax Target RR - Don Mario 5.9i, with row spacing of 0.50 m and 13 plants m-1, totaling 260,000 plants ha-1. For georeferencing of the area of study and sampling points was used a couple of topographic ProMarkTM3 receptors, making a relative positioning to obtain the georeferenced coordinates. To collect data (chemical analyzes of soil and crop yield) were sampled 10 blocks in the experimental area, each with an area of 20 m2 (20 meters long x 1 meter wide) containing two spaced adjacent rows of 0.5 m. Each block was divided into 20 portions of 1 m2, and from each were collected four subsamples at a distance of 0.5 m in relation to the lines of blocks, making up a sample depth for 0-10 cm a sample to 10-20 cm for each plot, totaling 200 samples for each depth. The soybean crop was performed on the blocks depending on maturity, and in each block was considered a bundle at each meter. In the data analysis, it was performed a diagnosis of multicollinearity, and subsequently a path analysis of the main variables according to the explanatory variables (range of chemical attributes: pH, K, P, Ca, etc.). The results obtained by the path analysis of the parameters of the semivariogram of soil chemical properties, indicated that only the Fe, Mg, Mn, organic matter (OM), P and Saturation by bases (SB) exerted direct and indirect effects on soybean productivity, although they have not presented spatial variability, indicating that the distribution of blocks in the area was unable to identify the spatial dependence of these elements, making it impossible to draw up maps of the chemical attributes for use in PA.

Detecção de região saliente em imagens usando dissimilaridade de cor e amostragem por pixeis aleatórios

The purpose of this work is to demonstrate and to assess a simple algorithm for automatic estimation of the most salient region in an image, that have possible application in computer vision. The algorithm uses the connection between color dissimilarities in the image and the image’s most salient region. The algorithm also avoids using image priors. Pixel dissimilarity is an informal function of the distance of a specific pixel’s color to other pixels’ colors in an image. We examine the relation between pixel color dissimilarity and salient region detection on the MSRA1K image dataset. We propose a simple algorithm for salient region detection through random pixel color dissimilarity. We define dissimilarity by accumulating the distance between each pixel and a sample of n other random pixels, in the CIELAB color space. An important result is that random dissimilarity between each pixel and just another pixel (n = 1) is enough to create adequate saliency maps when combined with median filter, with competitive average performance if compared with other related methods in the saliency detection research field. The assessment was performed by means of precision-recall curves. This idea is inspired on the human attention mechanism that is able to choose few specific regions to focus on, a biological system that the computer vision community aims to emulate. We also review some of the history on this topic of selective attention.