Show simple item record

dc.creatorCao, Hong Thang 1976-
dc.date.accessioned2018-04-26T15:39:06Z
dc.date.available2018-04-26T15:39:06Z
dc.date.created2018-06
dc.date.issued2018-04-26
dc.date.submittedJune 2018
dc.identifier.urihttp://hdl.handle.net/10388/8526
dc.description.abstractTo meet the high demand for supporting and accelerating progress in the breeding of novel traits, plant scientists and breeders have to measure a large number of plants and their characteristics accurately. A variety of imaging methodologies are being deployed to acquire data for quantitative studies of complex traits. When applied to a large number of plants such as canola plants, however, a complete three-dimensional (3D) model is time-consuming and expensive for high-throughput phenotyping with an enormous amount of data. In some contexts, a full rebuild of entire plants may not be necessary. In recent years, many 3D plan phenotyping techniques with high cost and large-scale facilities have been introduced to extract plant phenotypic traits, but these applications may be affected by limited research budgets and cross environments. This thesis proposed a low-cost depth and high-throughput phenotyping mobile platform to measure canola plant traits in cross environments. Methods included detecting and counting canola branches and seedpods, monitoring canola growth stages, and fusing color images to improve images resolution and achieve higher accuracy. Canola plant traits were examined in both controlled environment and field scenarios. These methodologies were enhanced by different imaging techniques. Results revealed that this phenotyping mobile platform can be used to investigate canola plant traits in cross environments with high accuracy. The results also show that algorithms for counting canola branches and seedpods enable crop researchers to analyze the relationship between canola genotypes and phenotypes and estimate crop yields. In addition to counting algorithms, fusing techniques can be helpful for plant breeders with more comfortable access plant characteristics by improving the definition and resolution of color images. These findings add value to the automation, low-cost depth and high-throughput phenotyping for canola plants. These findings also contribute a novel multi-focus image fusion that exhibits a competitive performance with outperforms some other state-of-the-art methods based on the visual saliency maps and gradient domain fast guided filter. This proposed platform and counting algorithms can be applied to not only canola plants but also other closely related species. The proposed fusing technique can be extended to other fields, such as remote sensing and medical image fusion.
dc.format.mimetypeapplication/pdf
dc.subject3D, plant phenotyping, image processing, counting canola branches & seedpods, multi-focus fusion image technique, low-cost depth camera, Argos3D P100
dc.titleA Low-cost Depth Imaging Mobile Platform for Canola Phenotyping
dc.typeThesis
dc.date.updated2018-04-26T15:39:06Z
thesis.degree.departmentElectrical and Computer Engineering
thesis.degree.disciplineElectrical Engineering
thesis.degree.grantorUniversity of Saskatchewan
thesis.degree.levelMasters
thesis.degree.nameMaster of Science (M.Sc.)
dc.type.materialtext
dc.contributor.committeeMember Karki, Rajesh 
dc.contributor.committeeMemberStavness, Ian
dc.contributor.committeeMemberChen, Li
dc.contributor.committeeMemberDinh, Anh
dc.creator.orcid0000-0002-5122-443X


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record