Index for the Evaluation of Segmentation (IAVAS): An Application to Agriculture

Abstract

The use of segmentation algorithms for determining the boundaries of cultivated plots is an essential step in the process of agricultural land use classification using digital remote sensing imagery. Thus, the main objective of the present research was the development of a quantitative method (IAVAS) for evaluating the results of segmentation of space imagery. The proposed methodology defines criteria for selecting thresholds (area and similarity) for the segmentation algorithm (region-growing approach). From the results obtained it was verified that the quantitative methodology proposed provides a suitable and efficient way to identify the best segmentation thresholds for agricultural land use classification.