Search Filter
- Publications
- Abstract of Theses and Dissertations
- Database
- Development of a Laboratory Scale Sorting Machine for Mandarin Orange...
Development of a Laboratory Scale Sorting Machine for Mandarin Orange (Citrus reticulata Blanco) Using Machine Vision
Thesis Abstract:
Citrus is one of the most important fruits in Myanmar and its annual production reached 353,050 metric tons in 2015. From 2009 to 2015, Myanmar’s average growth rate for the citrus production is three percent per year. In this study, a laboratory scale sorting machine for mandarin orange was developed. The system consisted of a belt conveyor, a webcam located in a lighting chamber, a feeding channel with receiving tray, and a computer with MSS software.
Results of statistic test showed that the orientation of mandarin orange with respect to the camera did not affect the software’s ability to analyze the images correctly. The software achieved 93.25 percent in sorting accuracy. In addition, the software evaluated mandarin orange’s equatorial diameter at 98.81 percent accuracy, while sphericity was evaluated with an accuracy of 97 percent. The processing time was 0.13 second per fruit.
During dynamic calibration, average system accuracies in terms of classifying mandarin oranges at different conveyor speeds, 10.25 cms-1, 8.36 cms-1, and 7.50 cms-1 were 92.67 percent, 92.67 percent, and 90.67 percent, respectively. The capacity of the sorting machine prototype was 1,161 fruits per hour at the highest conveyor speed.