An iterative region growing algorithm to generate fuzzy management zones within fields
Résumé
Management zones are within-field spatial units that are more homogeneous than the rest of the field and that have different factors or values than other areas of the field. While most algorithms tend to delineate fixed management zone borders, it is acknowledged that within-field variations between management zones are more gradual than sharp, mainly because of smooth variations in soil and landscape characteristics. In this work, an approach is proposed to account for the transient observations between management zones by delineating within-field fuzzy management zones. Observations within transition zones could belong to more than one zone, and this perhaps should be accounted for during the delineation process. Theoretical fertilization-based case studies, i.e. high and weak environmental constraints, were considered to show the interest of membership frequencies to refine yield-based management zones when user's expertise and management decisions come into play.