Automatic Object Painting with SCARA Robot Using Computer Vision

Abstract

Recognizing and accurately classifying colors in industrial applications is a critical challenge in developing robotics painting applications. To achieve this, many painting robots are attached with expensive color sensors. However, these sensors are coming with many drawbacks such as color ranges limitation and sensitivity to illumination in addition to their high costs. In the last decades, camera systems gained importance in robotics applications with the power presented by the computer vision techniques. The main objective of this paper is to design an automation line that includes a robot and camera system to perform painting in different colors, with various illumination conditions at cheaper costs. The proposed system can be used to paint multiple colors effectively and accurately. The power of the system comes from the color detection and classification algorithm that is designed using computer vision techniques. The algorithm is designed under C++ environment using OpenCV library. The system will able to detect all colors that are adjusted/predefined offline by the user and to work in different illumination conditions. The end-effector of the robot consists of two main parts, a camera to detect the desired color and an automatic spray gun to perform the painting operation. The proposed system detects will do so based on a small sticker pasted on the object that will be painted. When the desired color is detected, the system starts the painting operation. Moreover, the system has the capability to automatically cleaning the spray gun and the connected tubes in the case of having the successive objects to be painted in different colors.