• Infrared and Laser Engineering
  • Vol. 51, Issue 6, 20210446 (2022)
Pengfei Jia1, Quanzhou Liu1, Kai Peng2、*, Zhanqi Li1, Qipei Wang1, and Yiding Hua1
Author Affiliations
  • 1CATARC (Tianjin) Automotive Engineering Research Institute Co., Ltd., Tianjin 300300, China
  • 2School of Mechanical Engineering, Hebei University of Technology, Tianjin 300130, China
  • show less
    DOI: 10.3788/IRLA20210446 Cite this Article
    Pengfei Jia, Quanzhou Liu, Kai Peng, Zhanqi Li, Qipei Wang, Yiding Hua. Front vehicle detection based on multi-sensor information fusion[J]. Infrared and Laser Engineering, 2022, 51(6): 20210446 Copy Citation Text show less
    Network structure diagram of improved SSD algorithm
    Fig. 1. Network structure diagram of improved SSD algorithm
    Scale distribution graph of candidate boxes
    Fig. 2. Scale distribution graph of candidate boxes
    (a) Long focal length image; (b) Short focal length image; (c) Image with fusion processed; (d) Partial image without fusion processing; (e) Partial image with fusion processed
    Fig. 3. (a) Long focal length image; (b) Short focal length image; (c) Image with fusion processed; (d) Partial image without fusion processing; (e) Partial image with fusion processed
    (a) Collected vehicle pictures; (b) Improved SSD algorithm for vehicle detection result in complex environment
    Fig. 4. (a) Collected vehicle pictures; (b) Improved SSD algorithm for vehicle detection result in complex environment
    (a) Radar detection results in dynamic target simulation; (b) Influence of Lifetime parameters on radar detection results
    Fig. 5. (a) Radar detection results in dynamic target simulation; (b) Influence of Lifetime parameters on radar detection results
    Installation location of camera and radar on real vehicle
    Fig. 6. Installation location of camera and radar on real vehicle
    (a) Experimental picture; (b) Data collection
    Fig. 7. (a) Experimental picture; (b) Data collection
    (a) Data interaction between camera and radar;(b) Vehicle test results
    Fig. 8. (a) Data interaction between camera and radar;(b) Vehicle test results
    Calibration method Total sample size X direction image residual mean/pixel Residual variance of the X-direction image Y direction image residual mean/pixel Residual variance of the Y-direction image Total error per pixel
    Traditional joint calibration1 0000.23640.09820.28630.09920.3712
    Proposed method1 0000.18520.05680.19830.06130.2713
    Table 1. Image residuals and overall errors of calibration results
    WeatherVehicle number Detection accuracy of Ladar Detection accuracy of vision Detection accuracy of information fusion Fusion false detection rate
    Sunny332886.3%87.2%95.3%0.3%
    Cloudy189688.5%83.6%93.8%0.4%
    Night (Illumination)127589.2%80.4%91.7%0.6%
    Table 2. Comparison table of vehicle detection accuracy
    Pengfei Jia, Quanzhou Liu, Kai Peng, Zhanqi Li, Qipei Wang, Yiding Hua. Front vehicle detection based on multi-sensor information fusion[J]. Infrared and Laser Engineering, 2022, 51(6): 20210446
    Download Citation