Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Automatically Identifying floral contours and vascular bundles in 3D images

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  2018 ASABE Annual International Meeting  1800390.(doi:10.13031/aim.201800390)
Authors:   Yi-Hsiang Wang, Hao-Chun Hsu, Wen-Chieh Chou, Yan-Fu Kuo
Keywords:   Micro Computerized-tomography (micro-CT), 3D images, Hessian detector, surface images, vector superposition.

Abstract. Micro computerized-tomography (micro-CT) has started being used for studying structures and textures of biological samples in three-dimensional (3D). Because of the complex essence of 3D images, conventionally, the identification of traits in 3D images is performed manually. Manual identification is, however, time-consuming and can be subjective owing to researchers‘ experience. This work proposed to identify lobe contours and vascular bundles for flowers in genus Sinningia automatically. In the study, volumetric 3D flower images were acquired using micro-CT. Denoising algorithms were applied to improve the quality of the 3D images. Vascular bundles were then detected using Hessian detectors. The volumetric images were next converted to surface images. Lobe contours were then identified using vector superposition and the surface images. The experiment showed that the proposed methods can identify first-order vascular bundles and lobe contours at rates of 85.67% and 90.87%, respectively. The overall successful detection rate for detecting both the first-order veins and corolla contours was 81.71%.

(Download PDF)    (Export to EndNotes)