| Autor: |
Shahria, Md. Tanzil, Sunny, Md. Samiul Haque, Zarif, Md Ishrak Islam, Ahamed, Sheikh Iqbal, Rahman, Mohammad H. |
| Předmět: |
|
| Zdroj: |
IMEOM Conferences - Dhaka, Bangladesh; 2021, p698-705, 8p |
| Abstrakt: |
As many factors that can be attributed to a high-performance robotic system and with increasing demand for robotic application in automation or as assistive devices, research in this area has received much attention. Robot manipulation is one of the vital components of the industrial or assistive robotic system. The utilization of various vision-based approaches and different associated algorithms enhances the overall robot manipulation system's performance, functionality, and efficiency. Implementing a robust and accurate vision-based robot manipulation system is still today's challenge in robotics applications because of sensing and environmental uncertainties. In this paper, a computer vision-based object detection and localization system for robot manipulation system is proposed, which enables the robot to locate any object in the 3D space. The system uses a set of RGB camera, depth sensor, and vision processor that allows the system to get the coordinates of any objects in 3D space in terms of the camera to map and localize the object. Additionally, data associates are compared by statistical distance. Finally, a series of systematic experiments are performed to verify the reliability and accuracy of the proposed system. The system can successfully detect different objects and calculate data from the image frame with an accuracy of 94%. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
