Optical Techniques for Fruit Firmness Assessment
Rowe, P. (2015). Optical Techniques for Fruit Firmness Assessment (Thesis, Doctor of Philosophy (PhD)). University of Waikato, Hamilton, New Zealand. Retrieved from https://hdl.handle.net/10289/9208
Permanent Research Commons link: https://hdl.handle.net/10289/9208
This thesis describes the design and development of a new high-speed multispectral imaging (MSI) system compatible with a commercial grading line. The purpose of this system was to carry out spatially resolved spectroscopy to assess fruit firmness. Captured images were analysed using diffusion theory and modified Lorentzian models to extract a sample’s optical properties (absorption and reduced scattering coefficients) and optical parameters respectively. The high-speed MSI system was designed to capture images of fruit using a high-resolution complementary metal–oxide–semiconductor camera, 12.5 mm lens, and discrete lasers operating at 685, 850, and 904 nm. Each laser illuminates a separate fruit, and the camera captures the interacting light with a single frame encompassing all three fruit. Depending on the size of each fruit the spatial resolution with the 12.5 mm lens ranged from 0.15 to 0.22 mm/pixel. Initial measurements were made on 200 ‘Royal Gala’ apples to identify the relationships between the optical properties or parameters and either acoustic or the industry standard penetrometer firmness measurements. Performance of the high speed MSI system was poor compared to the results seen in the literature using alternative spatially resolved spectroscopic systems and other apple varieties. Only weak correlations (R = 0.33) were found between the individual optical measurements and firmness. Unsatisfactory performance from the high-speed system led to the development of a static MSI system to measure stationary fruit and the development of an inverse adding-doubling (IAD) system to provide an independent measurement of the samples optical properties. The purpose of these systems was to help understand the measurement, reduce variability, and give an indication of the upper level of performance possible. The static MSI system featured a number of improvements including the addition of a 980 nm laser, the elimination of an asymmetry caused by laser polarisation, improved temperature control, an electronic shutter system, precise location control of the fruit, and a new 25 mm lens improving spatial resolution (0.057mm/pixel). A second study was carried out using the new MSI and IAD systems on 92 ‘Royal Gala’ apples. Fruit were sliced to expose a flat measurement surface eliminating variation caused by fruit curvature and skin pigments. With these refinements and simplifications the relationships between optical properties or parameters and penetrometer firmness strengthened. As fruit softened and penetrometer firmness fell the reduced scattering coefficient measured by both the IAD and MSI system increased with correlation coefficients ranging from -0.62 to -0.70. The absorption coefficients measured by the two systems showed the expected features related to the absorption of chlorophyll and carotenoid pigments, and water absorption. As the fruit softened chlorophyll absorption decreased as the pigments are broken down and carotenoid absorption increased as new pigments are synthesised. No useful relationships were identified between the optical measurements and acoustic firmness. Multiple linear regression models were formed to predict penetrometer firmness using either the optical properties or modified Lorentzian parameters. The best performing model used a combination of the absorption and scattering coefficients, and had a correlation coefficient of 0.8 and a standard error of 5.87 N.
University of Waikato
- Higher Degree Theses