Image segmentation algorithms for corn kernel images with touching kernel regions
2006
Namuco, L.A.S.
A computer vision system for classification and evaluation of individual corn kernels will result in a more efficient, more reliable, and cheaper physical analysis system for corn graders. One problem of such system would be the identification of the individual corn kernels. Taking images of a single corn kernel at a time will make solving this problem easier but is tedious and expensive. A good image acquisition scheme for such a problem would be one where the kernel samples are simply laid out in bulk and an image taken. However, segmentation of individual corn kernels will be a problem as corn kernels will be touching and overlapping each other. Further identification of individual kernels will also prove to be difficult because of the many sizes and shapes of corn kernels. This study presented image segmentation schemes for identification of individual kernels from images of corn kernels taken in bulk. Distance transformations and modifications of the watershed algorithm were used for image segmentation. One algorithm and two distance transformation schemes were identified which were able to produce the best image segmentation results.
Afficher plus [+] Moins [-]Mots clés AGROVOC
Informations bibliographiques
Cette notice bibliographique a été fournie par University of the Philippines at Los Baños
Découvrez la collection de ce fournisseur de données dans AGRIS