Identification and Sorting of Objects by Color using Robotic Arm

A novel color-based object sorting system that uses a robotic arm controlled by a Convolutional Neural Network (CNN) model for efficient real-time classification and sorting. The system initially uses a USB camera to capture images of objects in RGB format. To enhance accuracy under various lighting conditions, these images are converted to HSV format, allowing separation of color (hue) from luminance. This transformation enhances the consistency and accuracy of color-based classification, especially under variable lighting conditions. The CNN model architecture is trained on a dataset of primary colors (red, green, and blue) plus yellow for more variety, and consists of convolutional and pooling layers for feature extraction and dense layers for classification into four color classes. This system demonstrates a practical approach to real-time automated sorting applications, achieving effective segmentation, classification, and handling of multiple object types through color recognition.