Digital Image Processing Algorithms
for Banana (Musa acuminata ‘Lacatan’) Size Classification
Marvin C. Petingco1 and Pepito M. Bato2
ABSTRACT
A machine vision system (MVS) was developed to grade bananas of Lakatan variety. The system’s components included a SONY DCR-TRV 460 video camera, a TV Capture card, a 2.66 Ghz Pentium 4 personal computer and a lighting chamber. A customized computer program called Bg06 was written in Microsoft Visual Basic 6 and was used as the judgment software component. It was able to calculate the length and the diameter of the banana finger and used these parameters as classification indices to evaluate its size. At a camera height of 40 cm from the sample platform, the MVS yielded overall size grading accuracies of 93% and 92.31% based on estimated weight and extracted projected area, respectively. The software’s processing time per banana finger was 7.73 milliseconds giving a grading rate of approximately 465, 000 bananas per hour.
Keywords: machine vision system, MVS, Bg06, banana size, size classification
1 Instructor and 2 Associate Professor, Agricultural Machinery Division (AMD), Institute of Agricultural Engineering (IAE), College of Engineering and Agro-Industrial Technology (CEAT), University of the Philippines Los Baños (UPLB), 4031 College, Laguna, Philippines.