Image segmentation algorithms for corn kernel images with touching kernels regions
2007
Namerco, L.A.S. | Albacea, E. | Khan, C. | Samaniego, J., Philippines Univ. Los Banos, College, Laguna (Philippines). Inst. of Computer Science
Yellow corn is one of the most important crops in the Philippines. It is used as food, animal feed, and raw product for many manufactured goods such as corn oil, corn starch, and snacks. In the Philippines, corn grading is done by the Philippine Grain Standards, where quality parameters are determined by visual inspection and manual separation. This manner of analysis is subjective and depends on the skill, physical condition, bias of the classifier, and working conditions such as lighting. It is also very tedious, slow, costly, and prone to errors. A computer vision system for classification and evaluation of individual kernels will result in more efficient, more reliable, and cheaper physical analysis system. The objective of the result of the study was to determine image segmentation schemes for identification of individual kernels for images of corn kernels taken in bulk. Several image acquisition schemes for kernel samples were done and appropriate algorithms for kernel segmentation were developed for the image acquisition schemes. These image acquisition schemes and segmentation algorithms were then assessed in terms of their running time, accuracy, and ease of use. Distance transformations and modifications of the watershed algorithm were used for image segmentation. One algorithm and two distance transformation schemes that produced the best image segmentation were identified.
显示更多 [+] 显示较少 [-]