Electronic Image Stabilizing Algorithm for Image Tracking System on Vehicle

罗诗途,王艳玲,张玘,罗飞路
DOI: https://doi.org/10.3321/j.issn:1004-924x.2005.01.016
2005-01-01
Optics and Precision Engineering
Abstract:A new image stabilization method is presented. First, an effective registration strategy is used which is from coarse to precious, and from local matching to global registration. In the coarse registration part, the improved SSDA algorithm is presented based on projection mean, and applied in fast template matching; for the precious registration, the improved regional counting method based on joint-histogram is applied to image sequence on vehicle, getting out of heavy floating-point operation while precision is guaranteed. Then, the adaptive mean motion filtering is introduced, which adaptively selects the size of moving window not only to smooth the motion curve, but also to prevent over-stabilization. The experimental results show that this algorithm can realize image registration on vehicle accurately and quickly, with the registration precision up to “sub-pixel” level, and can satisfy the demand of real-time and precision for image tracking system on vehicle.
What problem does this paper attempt to address?