Tomographic Image Segmentation Model for Features Extraction of Oilseeds Based on Graph Theory

This paper presents an evaluation model of oilseeds based on both the use of X-ray computed tomography, concepts of computational semantics, and digital image processing techniques. It was prepared to segment and extract characteristics of oilseeds digital images. In this context, we have applied the segmentation method based on Graph Theory and taking into account the relationship between the pixels belonging to the object and the background of the image. Besides, geometric descriptors have been included to allow an enrichment of the information obtained. Results have shown the useful of the model for the selection of oilseeds aiming to gains in productivity and quality.