• Acta Photonica Sinica
  • Vol. 50, Issue 10, 1011002 (2021)
Shuai YUAN, Xiang YAN, Jingxian XU, Wenrui ZHU, and Hanlin QIN*
Author Affiliations
  • School of Physics and Optoelectronic Engineering,Xidian University,Xi'an 710071,China
  • show less
    DOI: 10.3788/gzxb20215010.1011002 Cite this Article
    Shuai YUAN, Xiang YAN, Jingxian XU, Wenrui ZHU, Hanlin QIN. Cognitive Imaging Lidar Based on Deep Learning(Invited)[J]. Acta Photonica Sinica, 2021, 50(10): 1011002 Copy Citation Text show less
    Operating principle of cognitive imaging lidar based on deep learning
    Fig. 1. Operating principle of cognitive imaging lidar based on deep learning
    Block diagram of module internal structure
    Fig. 2. Block diagram of module internal structure
    Comparison of object detection results under different laser power in the same scene
    Fig. 3. Comparison of object detection results under different laser power in the same scene
    Scanning structure and scanning field of view control model
    Fig. 4. Scanning structure and scanning field of view control model
    Object detection results under different scanning resolutions in the same scene
    Fig. 5. Object detection results under different scanning resolutions in the same scene
    Flow chart of cognitive feedback for deep learning
    Fig. 6. Flow chart of cognitive feedback for deep learning
    Adaptive adjustment strategy of cognitive scanning field of view for deep learning
    Fig. 7. Adaptive adjustment strategy of cognitive scanning field of view for deep learning
    Comparison of conventional model detection result and cognitive feedback detection result
    Fig. 8. Comparison of conventional model detection result and cognitive feedback detection result
    Comparison of target detection effects between conventional model and cognitive feedback in remote multi-object scene
    Fig. 9. Comparison of target detection effects between conventional model and cognitive feedback in remote multi-object scene
    Comparison of target detection effects between traditional mode and cognitive feedback in long-range target squint scene
    Fig. 10. Comparison of target detection effects between traditional mode and cognitive feedback in long-range target squint scene
    WavelengthPulse durationPulse repetition frequencyAverage powerPeak powerDivergenceSpot diameter
    1.55 μm10 ns50 kHz25 W1 kW0.351 mrad12.2 mm
    Table 1. Laser source parameters
    MethodRecall/%Precision/%
    ConventionalCognitiveConventionalCognitive
    SECOND71.6276.5854.8865.23
    PointPillars72.9377.3155.4665.88
    CenterPoint75.2180.0254.3164.67
    Table 2. Comparison of recall and precision of car targets under traditional mode and cognitive mode
    Shuai YUAN, Xiang YAN, Jingxian XU, Wenrui ZHU, Hanlin QIN. Cognitive Imaging Lidar Based on Deep Learning(Invited)[J]. Acta Photonica Sinica, 2021, 50(10): 1011002
    Download Citation