Search

O objetivo deste trabalho foi comparar dois métodos de levantamentos de solos. O primeiro foi realizado por meio de metodologias tradicionalmente utilizadas na distinção das unidades de paisagem e discriminação das classes de solos, como fotointerpretação e relação solo-paisagem. O segundo foi baseado na distinção das classes desolos por meio da sua resposta espectral, obtida em nível orbital. Para o estabelecimento dos atributos dos solos e sua classificação, as amostras de solo foram coletadas em duas profundidades, num grid com pontos distanciados de 500 m. Com essas amostras foram realizadas análises físicas e químicas dos solos. Nos locais de amostragem foi determinada, a partir da imagem orbital, a reflectância aparente do solo e, por meio de análise de agrupamentos, foram estabelecidas as unidades de paisagem. A fidedignidade de semelhança entre as unidades de paisagem obtidas pelos dois métodos foi avaliada pelo índice Kappa,cujo valor foi de 0,43, indicando que ambos os métodos produziram resultados semelhantes na discriminação de classes de solo.The objective of this study was compare two soil survey methods. Thefirst was performed by methods traditionally used to distinguish landscape units and soil class discrimination. The second was based on soil class distinction through orbital spectral response. In order to establish soil characteristics and their classification, soil samples were collected at two depths in a grid system, with a distance of 500 meters between points. With these samples, physical and chemical analyses were carried out. In the sampling points, theapparent reflectance of the soil, from the orbital image, was determined and, through cluster analysis landscape units were established. In order to evaluate the resemblance reliability between the landscape units established in each method, the Kappa index wasused, the value set for the confusion matrix was 0.43, indicating high quality in the comparison, showing that the non-conventional method was as close as the one carried out by photointerpretation.

The objective of this study was compare two soil survey methods. The first was performed by methods traditionally used to distinguish landscape units and soil class discrimination. The second was based on soil class distinction through orbital spectral response. In order to establish soil characteristics and their classification, soil samples were collected at two depths in a grid system, with a distance of 500 meters between points. With these samples, physical and chemical analyses were carried out. In the sampling points, the apparent reflectance of the soil, from the orbital image, was determined and, through cluster analysis landscape units were established. In order to evaluate the resemblance reliability between the landscape units established in each method, the Kappa index was used, the value set for the confusion matrix was 0.43, indicating high quality in the comparison, showing that the non-conventional method was as close as the one carried out by photointerpretation. [ABSTRACT FROM AUTHOR]