Explore UCD

UCD Home >

Research Projects in 2019-2020

Monday, 20 January, 2020

Classification of different brands of cheddar cheese using hyperspectral imaging (PhD)

Lei T and Da-Wen Sun

Sponsors: CSC-UCD Scholarship Scheme

The present study aimed to investigate the feasibility of application of NIR-HSI to classification brands of Cheddar cheese. For the hyperspectral data, a probability based PLSDA was used to discriminate 4 brands of Cheddar cheeses. 18 PLS components were selected and the correct classification rate was 85% for both the cross-validation set and testing set. It is hoped that this study should provide a theoretical foundation on discriminating cheese products and set directions for the broader application of HSI in relevant research field.

Non-destructive investigation on moisture content uniformity and shrinkage rate caused by microstructure effect during microwave-vacuum drying (PhD)

Lin XH and Da-Wen Sun

Sponsors: CSC-UCD Scholarship Scheme

Microwave-vacuum drying (MVD) are relatively time saving and efficient drying methods. Although the shrinkage between hot air drying and MVD have been compared, the shrinkage and moisture distribution between the fiber perpendicular and parallel samples, the samples with and without peel during MVD were still not investigated. Thus, this study applied hyperspectral imaging and computer vision to investigate the shrinkage and moisture distribution difference between the different ginger samples. The microwave-vacuum dryer used in the experiment was designed by Food Refrigeration and Computerized Food Technology (FRCRT, Dublin, Ireland). A laboratory hyperspectral imaging system was used to acquire the hyperspectral images of ginger sample in the reflectance model. A computer vision system was applied to acquire the images of samples during MVD. The quantitative models between the moisture content and the spectral data of the ginger slices based on diverse spectral pre-processing methods covering the raw spectra and those modified by standard normal variate (SNV) and multiplicative scatter correction (MSC) were established by partial least regression model (PLS). All the data analysis was implemented in Matlab R2016a (The MathWorks, Inc., Natick, Massachusetts, USA). The PLS models with SNV pretreatment was the optimal model to predict moisture content of ginger slices, and thus it was used to predict the moisture distribution of ginger slices. The shrinkage rate of the sample with peel were significantly higher than the sample without peel. The shrinkage rate of width and length of perpendicular samples is higher than parallel samples. In addition, the shrinkage rate of width and length in parallel samples were different. The shrinkage rate of width was higher than that of the length. In the above, the shrinkage rate and moisture distribution was different between perpendicular and parallel samples, and between samples with and without peel. The current study could provide an insight in to the moisture distribution and shrinkage difference between different cutting direction samples, and provide a direction for the improvement of drying food quality.

Comparison of hot water, microwave-assisted and deep eutectic solvents extraction of polysaccharides from t. Fuciformis and its water holding ability (PhD)

Chen TJ, Miao S and Da-Wen Sun

Sponsors: CSC-UCD Scholarship Scheme

This experiment was carried out to investigate the most efficient and eco-friendly polysaccharides extraction method of T. fuciformis and Polysaccharides of T. fuciformis, which will be extracted by using hot water, microwave-assisted, deep eutectic solvents. The highest yield level will be microwave-assisted. Deep eutectic solvents show a good and non-destructive performance in extraction as a green solvent. Water holding capacity of T. fuciformis’ Polysaccharides will be tested with casein. UV-visible spectrophotometric analysis will be applied to analyze the composition of Polysaccharides.The study will offer comparisons among T. fuciformis polysaccharide extraction methods and their water holding capacity with casein.

Diagnostics of reactive species generated by cold and hot plasma (PhD)

Zhang KX, Tiwari BK, Da-Wen Sun and Zhao M

Sponsors: CSC-UCD Scholarship Scheme

Introduction: Plasma is the fourth state of matter, consisting of highly excited species. Plasma treatment has been regarded as a novel process technique and been applied to a wide range products in food industry. Plasma generated reactive species (mainly refer to reactive nitrogen and oxygen species) play an important role in microbial decontamination, toxin degradation and surface modification of packaging materials.

Objective: This study aims to detect plasma generated reactive species using both wet chemical based colorimetric method and non-destructive optical absorption spectroscopy.

Methods employed: Deionized water (25 ml) was treated using cold and hot plasma for 5min, 10min and 20min, respectively. After the treatment, the concentration levels of H2O2, NO2-, and NO3- in the plasma activated water were detected using titanium oxysulfate (Sigma-Aldrich), Griess reagent (Sigma-Aldrich) and nitrate test kit (Merck), respectively. The optical absorption spectra (OAS) of plasma generated reactive species were collected using a CCS spectroscopy (CCS200, THORLABS) in the wavelength range of 350-900 nm with the illumination enhancement of a tungsten halogen lamp (QTH10/M, THORLABS). The CCS spectroscopy probe was placed vertically underneath the plasma nozzle and its position to tungsten halogen lamps was adjusted to achieve a high signal to noise ratio of the spectra. Absorption spectral data of each measurement was averaged from 50 times of scans.

Main result: Results of the chemical analysis revealed that the concentration levels of H2O2, NO2- and NO3- in the activated water were increased during the treatment time. Spectral results showed that spectral shape differences are observed between the air base (control) and plasma. Absorption spectra, which were acquired after 5, 10 and 20 min plasma treatments, lay out in a pattern following the logical order of treatment time.

Conclusion sentences: In conclusion, different concentration levels of reactive species under different treatment time were found in both of the chemical references and OAS spectral data. The future work will be investigated to explore the correlations between the chemical references and OAS spectral information.

Research on simultaneous and sequential ultrasound- and microwave-assisted methods for aqueous extraction of bioactive compounds from coffee silverskin (PhD)

Wen L, Da-Wen Sun and Tiwari BK

Sponsors: CSC-UCD Scholarship Scheme

U-MAE and M-UAE) were both applied, and the effects of different conditions were analysed for the extraction of phenolics from coffee silverskin. In addition, the combined extraction technologies were compared with their individual techniques (UAE alone and MAE alone) for the effects on the extraction yield, antioxidant capacity and contents of caffeine and some widely-known bioactive compounds from coffee silverskin.

S-UMAE achieved 25.9%, 16.7% 13.1% and 12.8% higher extraction yield than UAE, MAE, U-MAE and M-UAE, respectively. The antioxidant capacity and chlorogenic acids (CGAs) content were also enhanced significantly by S-UMAE than by the other methods. Caffeine content showed no significant (P > 0.05) difference between the extraction methods. Compared with U-MAE, extracts obtained by M-UAE exhibited higher antioxidant capacity (28.57% and 20.60% on DPPH and FRAP values, respectively) and 68.19% higher total content of CGAs. Three-cycle S-UMAE (20 min each cycle) under the optimised conditions achieved a similar yield of total phenolics obtained by three-cycle conventional methanol extraction (24 hours extraction each cycle). The micro-structure investigation of CSS showed the combination of UAE and MAE can be a more effective extraction technique for extraction of bioactive compounds from plant materials. This work shows the potential and compatibility of the application of combined extraction technologies on the extraction of bioactive compounds from CSS. It could be an effective option to accomplish a balance between extraction efficiency, product quality, consumption of solvent and production costs in the industry.

NIR hyperspectral imaging in visualizing moisture distribution of apple slices during hot air oven drying (MEngSc)

Valgadde PS, Da-Wen Sun and Lin XH

Sponsors: University College Dublin

NIR Hyperspectral imaging system is a robust and innovative technique which has proven to be an effective tool to detect moisture distribution in the sample. This is an accurate way of analysis which helps in providing superior quality food. Hyperspectral imaging is the combination of spectral and spatial information. Apples will be washed, peeled and sliced into infinite slabs of thickness 2,4 and 6 mm and 45-55mm diameter using a stainless-steel food slicer. A total 30 samples of the three mentioned thickness will then go for dehydration. Drying will be carried out in hot-air oven at 70℃ and moisture content will be checked every 30 min which will be recorded until the final moisture content reaches 22%. The dehydrated apple slices will be subjected to hyperspectral imaging system to visualize moisture distribution throughout the samples.

Hyperspectral imaging to measure the mosture content of kiwi slices using hot-air drying (PhD)

Yin YL, Da-Wen Sun and Lin XH

Sponsors: University College Dublin

Hyperspectral imaging is a rapid and non-destructive detection technique for quality analysis such as moisture content, which is operated by simultaneously offering spectral data. The purpose of this study was to demonstrate the feasibility of hyperspectral imaging in measuring moisture content of kiwifruit under hot-air drying process. The data processing and analysis were based on Matlab 2015a. Calibration model was based on partial least square regression (PLSR) to establish the relationship between spectral data and moisture distribution.

UCD Food Refrigeration & Computerised Food Technology

Food Refrigeration and Computerized Food Technology University College Dublin Agriculture & Food Science Centre Belfield, Dublin 4, Ireland.
T: +353 1 716 7342