Embedded Files
This work, which was performed by two senior ECEE students from The University of Memphis (Allan Abraham and Kameshwaran Balachandran). In this outreach undergraduate research project, we have presented a low-cost method to distinguish the quality of different olive oils. The proposed method is based on an indirect measurement of the chlorophyll molecules present when a green laser diode illuminates the oil sample. Oil blends can be classified into five classes (no olive oil, light olive oil, medium olive oil, olive oil, and extra virgin olive oil) by quantifying the ratio of the red channel versus the green channel along the laser illumination path from a color image. This undergraduate research project introduces students to an interdisciplinary application requiring the combination of optical spectroscopy (i.e., multicolor imaging), and image processing. In addition, due to the simplicity of the optical apparatus and computational analysis, high school students could implement and validate their own cost-effective oil-quality classification device.
Multicolor Imaging of Olive Oil Blends
When oil blends are excited by a green illumination beam, the chlorophyll molecules emit light within the red wavelength. Vegetable oils contain varying amounts of chlorophyll molecules and emit light at different color and intensity/power. The higher the number of chlorophyll molecules, the higher the quality of the olive oil blend and the more intense the red light is perceived. We propose to classify the olive oil by indirectly measuring the amount of chlorophyll molecules. Note that the olive-oil molecules are fluorescent molecules, emitting signals within the red wavelength spectrum when they are illuminated with green light. The below figure shows the experimental RGB images of four oil blends. For each RGB image, we have decomposed the color image into the three different channels. The below figure shows the corresponding green and red channels for each color image. Clearly, the content on both green and red channels changes for different oil blends. In particular, the red channel has more information when the concentration of olive oil particles in the oil blend is higher. We have estimated the information on both green and red channels by normalizing the mean intensity values within the laser illumination area for each channel. These values are shown in the below figure.
When oil blends are excited by a green illumination beam, the chlorophyll molecules emit light within the red wavelength. Vegetable oils contain varying amounts of chlorophyll molecules and emit light at different color and intensity/power. The higher the number of chlorophyll molecules, the higher the quality of the olive oil blend and the more intense the red light is perceived. We propose to classify the olive oil by indirectly measuring the amount of chlorophyll molecules. Note that the olive-oil molecules are fluorescent molecules, emitting signals within the red wavelength spectrum when they are illuminated with green light. The below figure shows the experimental RGB images of four oil blends. For each RGB image, we have decomposed the color image into the three different channels. The below figure shows the corresponding green and red channels for each color image. Clearly, the content on both green and red channels changes for different oil blends. In particular, the red channel has more information when the concentration of olive oil particles in the oil blend is higher. We have estimated the information on both green and red channels by normalizing the mean intensity values within the laser illumination area for each channel. These values are shown in the below figure.
Classification algorithm based on image processing
As a classifier function to discriminate the quality of the oil blend, we have used the ratio of the normalized mean intensity values between the red and green channels (e.g., R/G). The below figure plots this ratio for the fifty tested oil blends. Based on this ratio, we have classified oil blends as: 'No Olive Oil' if R/G < 1; 'Light Olive Oil' if 1 ≤ R/G < 1.5; 'Medium Olive Oil' if 1.5 ≤ R/G < 2; 'Olive Oil if 2 ≤ R/G < 2.5; and 'Extra Virgin Olive Oil' if R/G ≥ 2.5.
As a classifier function to discriminate the quality of the oil blend, we have used the ratio of the normalized mean intensity values between the red and green channels (e.g., R/G). The below figure plots this ratio for the fifty tested oil blends. Based on this ratio, we have classified oil blends as: 'No Olive Oil' if R/G < 1; 'Light Olive Oil' if 1 ≤ R/G < 1.5; 'Medium Olive Oil' if 1.5 ≤ R/G < 2; 'Olive Oil if 2 ≤ R/G < 2.5; and 'Extra Virgin Olive Oil' if R/G ≥ 2.5.
Credits
Credits
- The code was developed in and Matlab R2023a. Natick, Massachusetts: The MathWorks Inc.
Citation
Citation
If using this dataset or code for publication, please kindly cite the following:
If using this dataset or code for publication, please kindly cite the following:
- A. Abraham, K. Balachandran, and A. Doblas, “Research Outreach Interdisciplinary Activity to classify olive oil blends integrating multicolor imaging,i mage processing, and machine learning,” Undergraduate Research Journal 3(2), 6. doi:/10.58361/2766-3590.1062
Support or Contact The Principal Investigator of this REU project is Dr. Ana Doblas (adoblas@umassd.edu).
Support or Contact The Principal Investigator of this REU project is Dr. Ana Doblas (adoblas@umassd.edu).
Page updated
Google Sites
Report abuse