Electrical Component Classification Using Deep Learning

This paper introduces a real-time CNN-based classification and detection solution for breadboards, LED (Light Emitting Diode), LCD (Liquid Crystal Display), Battery and Arduino as common electronics components. This system trains CNN using an image dataset. We propose a real-time detection framework for identifying and locating the components using OpenCV with laptop webcam and using Arduino IDE with ESP32 CAM on live video streams. The model reads frames from the webcam, identifies regions of the object and classifies a component associated with higher confidence score, while bounding boxes are drawn around detected objects in each frame. With a few minor misclassifications between meats that look like components, the system has great potential to be applied in electronics sorting and education for manufacturing automation. We plan to continue adding more data, improving the detection accuracy of the detector and integrating object tracking for better performance.