Abstract
Development a visual-guided autonomous arm robot for general working application in service workshop require some preliminary works/research to ensure the quality and reliability of robot mainly on object detection/recognition and object pose estimation. We have experimented robot vision for this purpose using Raspberry Pi and single web camera supported by Python-OpenCV programming using color-base and contour-base detection algorithm for object recognition and Triangulation similarity method for object pose estimation. Experiment results showed that color-base detection is 22% faster than contour-based object detection for colorful tooling object without disturbance same color from environment. However, contour-base detection is more effective for target working object detection than color-base. Light illumination and disturbance from environment should be managed for successful object detection. Triangulation linearity method is simple and fastest method for tooling object position estimation when tooling object is a known sized object. Experiment result showed error only 2% for distance estimation using this method compared with actual.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call