A systematic assessment of remote sensing approaches for agricultural zonation and supporting precision agriculture

Remote sensing (RS) technologies generate large amounts of data that can be highly valuable for supporting field zonation and precision agriculture. However, numerous processing approaches have been proposed, with no clear guidelines for their application. This study aims to assess the effects of (i) the vegetation index (VI), (ii) the clustering algorithm, and (iii) the timing of image acquisition on the quality and consistency of zonation results. Sentinel-2 images were analyzed over two growing seasons at two experimental sites in Northern Italy. Several ground observations were integrated to define the benchmark. The results indicated that the choice of VI and clustering method had only a minor effect on zonation, whereas image selection had a significant influence. The best results were obtained during the crop growing season and under dry conditions. Overall, this study demonstrates that systematic analyses can provide practical guidance for implementing RS technologies to support precision management practices.