Road vehicle detection based on multi-sensor fusion

Guoxing Wu,Chunxia Zhao,Jiayin Liu,bin Hu
DOI: https://doi.org/10.13245/j.hust.15S1060
2015-01-01
Abstract:To efficiently detect the vehicle on road,a vehicle detection method was proposed based on millimeter wave (MMW)radar and vision sensors.Fusion strategy was carried out on the underlying data.In radar data channel,Hough transform and Chebyshev's theorem were combined together to re-move the off-road radar data.Then the reversed data were regarded as targets and used to generate the sparse probability graph.In image channel,saliency map for the vehicle was generated by the visual saliency model,which was built upon the low-level image features set and sparse probability graph. The final vehicle detection was achieved by the over segmentation method.The comprehensive evalua-tions with the structural road scenarios demonstrate that the proposed method gives a superior per-formance than those pure vision based methods,which is sensitive to the illumination various.
What problem does this paper attempt to address?