3D Object Localization Using 2D Estimates for Computer Vision Applications
| Autor: | Muhammad Usman, Yawar Rehman, Malik Zohaib Nisar, Taha Hasan Masood Siddique, Tahira Rafiq, Muhammad Sohail Ibrahim |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Computer science
business.industry Distortion (optics) 3D reconstruction ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Process (computing) Location awareness Object (computer science) computer.software_genre Computer vision Artificial intelligence MATLAB business computer Pose computer.programming_language Camera resectioning |
| Zdroj: | 2021 Mohammad Ali Jinnah University International Conference on Computing (MAJICC). |
| DOI: | 10.1109/majicc53071.2021.9526270 |
| Popis: | Computer vision based applications have received notable attention globally due to the interaction with the physical world. In this paper, a novel method for object localization based on camera calibration and pose estimation is discussed. The 3-dimensional (3D) coordinates are computed by taking multiple 2-dimensional (2D) images from different view that meets the requirement in camera calibration process. Several number of steps are involved in camera calibration including estimation of intrinsic and extrinsic parameters for the removal of lens distortion, estimation of object’s size and camera location. Besides, a technique to estimate the 3D pose using 2D images is proposed and the results of camera parameters and localization are applied for the 3D reconstruction. The hardware implementation of the proposed approach is implemented on HP core i5 with the MATLAB support packages and experimental results are validated for both camera calibration and pose estimation. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|