Computer Vision-Based System for Classification and Sorting Color Objects

Abstract

Abstract Interest in computer vision has grown considerably in industrial applications of robotic arms such as packaging, sorting, and other applications. The real challenge is how to improve the current sorting system based on image processing in the four-points to identify, manipulate, select and sort objects depending on color and geometry. This paper presents a mechatronic sorting system with the application of image processing using a web camera by sensing existing parts and capturing them in real-time, then defining colors and analyzing geometric data. This information is processed for the handling mechanism. A robotic arm has 5 DOF is used in sorting the objects have different geometries (circle, rectangular, and triangle) and a different color (red, green, blue, brown, orange, and cyan). The project deals with an automated material handling system. It aims to classify color objects by color, size, and geometry, which comes on a conveyor belt by selecting and placing the parts in their pre-programmed location. Image processing is applied in EmguCV and visual studio software. An image is a matrix of pixels, image processing is the processes applied on image to convert it into digital form and implement some process on it to improve an image. The Canny edge detection algorithm is used for shape identification in image processing. A microcontroller with computer vision is used to automate the sorting process. Experimental results reveal the effectiveness, high accuracy and low cost of computer vision with a robotic arm in the sorting process according to color and shape.