Confused Pixels Interference in Maps of Agricultural Management Zones

Abstract

Management zones can be delimited using fuzzy logic, a technique that assigns values of degrees of pertinence to each pixel of a map. When the value tends to 1, these degrees indicate that there is certainty that the pixel belongs to a certain class of the management zone. However, in the boundary region between classes, degrees of pertinence do not tend to 1, indicating that there is confusion about which class such pixels belong. Depending on the area occupied by confused pixels, the use of management zones as a precision agriculture technique can be compromised. Thus, the behavior of the area occupied by pixels with different degrees of pertinence was evaluated as a function of the amount of information used to generate the management zones. Those zones were generated based on altitude, soil apparent electrical conductivity in soil depths of 0.20 m and 0.40 m, soil water content and clay content. When adding information to generate the management zones, there was an increase in the area occupied by pixels with degrees of pertinence lower than 0.5. However, the insertion of more than one layer of information to delineate the management zones improved the concordance between the management zones and the maps of the soil attributes. We suggest that some samples should be distributed in the border regions between the management zones, when these are delimited from the use of two or more variables.