A Fast Method to Estimate Sensor Translation

Abstract

An important problem in visual motion analysis is to determine the parameters of egomotion. We present a simple, fast method that computes the translational motion of a sensor that is generating a sequence of images. This procedure computes a scalar function from the optical flow field induced on the image plane due to the motion of the sensor and uses the norm of this function as an error measure. Appropriate values of the parameters used in the computation of the scalar function yield zero error; this observation is used to locate the Focus of Expansion which is directly related to the translational motion .

Cite

Text

Sundareswaran. "A Fast Method to Estimate Sensor Translation." European Conference on Computer Vision, 1992. doi:10.1007/3-540-55426-2_28

Markdown

[Sundareswaran. "A Fast Method to Estimate Sensor Translation." European Conference on Computer Vision, 1992.](https://mlanthology.org/eccv/1992/sundareswaran1992eccv-fast/) doi:10.1007/3-540-55426-2_28

BibTeX

@inproceedings{sundareswaran1992eccv-fast,
  title     = {{A Fast Method to Estimate Sensor Translation}},
  author    = {Sundareswaran, Venkataraman},
  booktitle = {European Conference on Computer Vision},
  year      = {1992},
  pages     = {253-257},
  doi       = {10.1007/3-540-55426-2_28},
  url       = {https://mlanthology.org/eccv/1992/sundareswaran1992eccv-fast/}
}