A method of 3D nondestructive detection for plant root in situ based on CBCT imaging

• Published in: 2014 7th International Conference on Biomedical Engineering and Informatics pp. 110 - 115
• Main Authors: Zhou, Xuecheng, Cao, Xia, Zhang, Changling, Yan, Hui, Li, Yihai, Luo, Xiwen
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.10.2014
• Subjects: Computed tomography; Cone-Beam CT; CT imaging; Image reconstruction; Image segmentation; in situ; nondestructive detection; plant root; Soil; Three-dimensional displays; X-ray imaging
• ISSN: 1948-2914; 1948-2922
• DOI: 10.1109/BMEI.2014.7002753

Root is an important organ for plant to uptake nutrient and water from the surrounding soil. And the capability of plant to obtain nutrient and water is closely relevant to its root distribution and architecture under ground. Soil is the source of water and nutrients for plant; however, it also obstructs the observation and measurement of the root system. A thorough investigation of root morphology and biology for plant root systems requires advanced observation and measurement methods to characterize the root architecture in situ. In this study, a three-dimensional (3-D) visualization and nondestructive detection technique using X-ray computed tomography (CT) based on a flat panel detector is presented. An experimental unit was developed to examine the technical feasibility. First, the CT was designed to capture the sectional images of plant roots growing in soil or other media. Second, a series of program processing was conducted for root CT images involving sequential image preprocessing, image segmentation, 3-D reconstruction and visualization, and extraction of root parameters. Finally, the parameters of the 3-D architecture of roots obtained from the computer detection were analyzed to compare them with those from manual measurements. The results show that (1) the reconstructive accuracy of volume approaches 87%, (2) the relative error of structural root length ranges from 3.27 to 10.82% with an average of 7.07%, (3) the relative error of root diameter at branch nodes ranges from 3.00 to 12.18% with an average of 8.06%, (4) the minimum diameter of roots available for detection is 0.2 mm, and (5) the relative error of root branch angle ranges from 3.74 to 21.15% with an average of 10.72%. Moreover, the nondestructive measurement method can be used to determine the structure of plant roots automatically, consecutively and dynamically without shifting or remolding the original form of root system in situ.