Image data collection, processing, storage, and their application in smartphone food analysis

  • Yunfeng Zhao

Student thesis: Doctoral ThesisDoctor of Philosophy


This thesis proposes a new food analysis solution for on-site operation using custom hardware and software and widely available smartphone technology. A coherent set of image data collection, processing, and persistence techniques are developed in this thesis as building blocks for the solution. A spectral nonuniform illumination correction and an illumination matching algorithm are proposed to minimise any effect due to nonuniformly distributed illumination and varied illumination colour with considering variations in image formation in different cameras. The thesis introduces a novel colour alignment technique that works with non-standard corresponding colour points and only needs a minimum of two corresponding colour points during colour matching. A high-performance camera response model that represents camera responses with only a single latent variable is proposed for more accurate and rapid camera calibration. A practical smartphone modularised colourimetric reader for on-site image analytical data collection is developed for data acquisition. A specific colour subtraction algorithm is also proposed that removes interference due to coloured solutions during image data collection. To improve data integrity and robustness, a novel two-tier hybrid data storage scheme is proposed. This combines a blockchain network with traditional non-relational databases for trustable, efficient, and interoperable data persistence.

Date of AwardJul 2022
Original languageEnglish
Awarding Institution
  • Queen's University Belfast
SponsorsEC/Horizon 2020 Marie Skłodowska-Curie actions
SupervisorKaren Rafferty (Supervisor), Christopher Elliott (Supervisor) & Huiyu Zhou (Supervisor)


  • Smartphone-based sensing
  • image analysis
  • food safety

Cite this