Development of an Optical Colour Recognition Application for Automated Resistor Value Identification Using a Camera

This project presents the design and development of an optical colour recognition application for automated resistor value identification using a camera. The system was developed to address the limitations of manual resistor reading, which is prone to errors due to colour similarity, poor lighting, and human interpretation. The application integrates computer vision and machine learning techniques within a Django web framework, enabling users to upload or capture resistor images for automated analysis. Image pre-processing was performed using OpenCV to resize, denoise, and convert images into HSV and LAB colour spaces, thereby enhancing the visibility of bands under varying conditions. The resistor color bands were detected and segmented using a pre-trained YOLOv8 object detection model, and accurate color classification was achieved using Euclidean distance in LAB space. The resistance values were then calculated programmatically with tolerance ranges after the identified colours were mapped to common resistor codes. The system was able to calculate resistor values like 56Ω ±5% in real-world trials, demonstrating high accuracy across a variety of lighting and orientation scenarios. The developed application demonstrated strong potential for use in diagnostics, education, and electronics manufacturing, and it proved to be dependable, and cost-effective.