Abstract
In this paper, we present an efficient visual simultaneous localization and mapping (SLAM) for a fisheye lens. A fisheye lens can obtain more information through a wide field of view (FoV) compared with a pinhole camera model. However, fisheye lenses has strong radial distortion, necessitating proper distortion correction. Pre-existing visual SLAMs for fisheye lenses have disadvantages in that either the correction processing speed is slow or the information of the input image is deformed in the correction process. We solve these problems by backprojecting the descriptor from a distortion-removed domain to the original distorted images using a parametric camera model. In addition, the multi-pinhole cube map model is employed to utilize the entire FoV without image distortion and obtain scale and rotation invariance. In the optimization process, more stable optimization is achieved using a new Jacobian calculation method suitable for a sphere model. The performance is evaluated in terms of the overall system processing time and error comparison between the camera pose and ground truth when creating the SLAM map. Experimental results indicate that our proposed method has a high processing speed while having a similar or better accuracy compared with another method.
Sign-in/Register to access full text options 
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 callMore From: Journal of Institute of Control, Robotics and Systems
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
