AcurÃcia e precisÃo do mapeamento da flora nativa e dos reflorestamento em Minas Gerais / Accuracy and precision of the thematic mapping of native flora and reforestation of Minas Gerais

AUTOR(ES)
DATA DE PUBLICAÇÃO

2008

RESUMO

This study aimed to evaluate the overall accuracy and precision of thematic maps. Specifically aimed to evaluate the accuracy and precision of the classification of Thematic Mapping of Native Flora and Forestry of Minas Gerais. Data were collected from the field around the reference state, using simple random sampling procedure in proportion with the support of the itineraries Method (Belart et al., 2001). Thus, the sampling units - polygons determined by points - were taken randomly along the roads, paved or not. To this end, selected roads were mandatory, and additional travel. The size of the reference sample was calculated according to the formula for scaling the size of the sample (Meyer, 1984). In Chapter 2, the number of polygons sampled by sub-basin followed the relationship between the area of each class in this particular sub-basin in relation to the total area of class in the state. In Chapter 3, the same sample was analyzed by both classes as a thematic basins. Upon obtaining the reference data in the field, they were compared to the mapping, and they assigned the values 0 or 1 for failure or success, respectively. Subsequently plates were made of confusion and calculated correlation coefficients for total and for classes that comprised. Was also carried out to determine the accuracy of correctly classified polygons, and generate confidence intervals for this variable. The confidence interval for the proportion of actual polygons correctly classified by mapping was 86.67% Proportion ≤ ≤ 87.87% Real and is considered satisfactory. The figures given by the coefficient of overall accuracy was 87.27% while 81.61% of the Kappa Tau and 86.29%. Only the Kappa value showed less than the accuracy set a priori (85%). The classes had themes that were more confusion in the classification and Eucalyptus Forest Rain. Among the basins, which showed more confusion between the classes that were of composed Mapping Rio Piracicaba / Jaguari, Pardo, and JucuruÃÃ Jequitinhonha. The proporcionalizaÃÃo obtained with the distribution of sample units according to the area that held the classes in thematic sub-basins, ensured that all classes have a minimum number of units to meet the pre-set error and generate confidence intervals for the proportion of correctly classified polygons. The use of roads as a basis for taking the sample was efficient since the road network of the area had great length and good spatial distribution, favoring the distribution of sampling units across the area.

ASSUNTO(S)

thematic mapping mapeamento temÃtico amostragem thematic mapping of native flora and reflorestation of minas gerais accuracy mapeamento da flora nativa e dos reflorestamentos de minas gerais acurÃria precisÃo manejo florestal precision sampling

ACESSO AO ARTIGO

Documentos Relacionados