4-band resistor recognition using LeNet-5

• Main Authors: Puno, John Carlo V., Rabano, Stephenn L., Velasco, Jessica S., Cabatuan, Melvin K., Sybingco, Edwin, Dadios, Elmer P.
• Format: text
• Published: Animo Repository 2022
• Subjects: Electric resistors; Neural networks (Computer science); Deep learning (Machine learning); Electrical and Computer Engineering
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/5749
• Institution: De La Salle University

Compound color object recognition application is a challenging problem. This problem is applied to the automatic reading of resistor values for 4-band resistors. The images of different resistor values are in a .jpeg extension. The readings are based on a standard resistor color-coding. In this paper, there are 4957 images in datasets with 420 categories. The study used the Lenet-5 algorithm to recognize 240 different resistor values of different wattage ratings, with either 5% or 10 % tolerance. The model is composed of 5 layers inclusive of the following: two 2D convolution layers, one flatten layer, and two dense layers. The test showed 99.6% accuracy with a test score of 0.0285 based on the 50-epoch training. Another test was done using additional flipped images, the model showed the test score is 0.0934 and a test accuracy of 99.2%.