• Infrared and Laser Engineering
  • Vol. 53, Issue 10, 20240217 (2024)
Ronghua LI1,2, Xinchen ZHOU1,2, Chuanxin WENG3, Haopeng XUE1,2..., Jinlong WU1,2 and Chenyu LIN1,2|Show fewer author(s)
Author Affiliations
  • 1Institute of Mechanical Engineering, Dalian Jiaotong University, Dalian 116028, China
  • 2Advanced Robot Sensing and Control Technology Innovation Center, Dalian 116028, China
  • 3Beijing Institute of Environmental Characteristics, Beijing 100854, China
  • show less
    DOI: 10.3788/IRLA20240217 Cite this Article
    Ronghua LI, Xinchen ZHOU, Chuanxin WENG, Haopeng XUE, Jinlong WU, Chenyu LIN. Non-contact infrared laser physical property inversion method for target surface based on SSA-GRNN[J]. Infrared and Laser Engineering, 2024, 53(10): 20240217 Copy Citation Text show less
    Flowchart of the physical inversion method
    Fig. 1. Flowchart of the physical inversion method
    Neural network model of GRNN
    Fig. 2. Neural network model of GRNN
    Physical inversion process
    Fig. 3. Physical inversion process
    Target to be tested
    Fig. 4. Target to be tested
    Echo intensity measurements of different materials at the same distance
    Fig. 5. Echo intensity measurements of different materials at the same distance
    Echo intensity measurement results at different distances with the same material
    Fig. 6. Echo intensity measurement results at different distances with the same material
    Measurement system for laser echo intensity (1-Laser driver; 2-Detector drive; 3-Target to be tested; 4-Detectors; 5-Infrared laser emitter; 6-Turntable; 7-Oscilloscope)
    Fig. 7. Measurement system for laser echo intensity (1-Laser driver; 2-Detector drive; 3-Target to be tested; 4-Detectors; 5-Infrared laser emitter; 6-Turntable; 7-Oscilloscope)
    Experimental program
    Fig. 8. Experimental program
    Comparison between output values and actual values of different materials
    Fig. 9. Comparison between output values and actual values of different materials
    Inversion probabilities of SSA-GRNN and GRNN prediction models at different distances
    Fig. 10. Inversion probabilities of SSA-GRNN and GRNN prediction models at different distances
    ParameterProperties
    SurfacePolyimideOriginalPaint
    SubstrateSteel& foilSteelSteel
    Reflection typeSpecularSemi specularDiffuse
    Table 1. Parameters of the target to be tested
    Evaluating indicatorEMAEMSERMS
    SSA-GRNNTraining set0.63151.76051.3268
    Test set0.30361.84691.3590
    GRNNTraining set13.6392548.399423.4179
    Test set3.4873147.839712.1589
    Table 2. Evaluation of laser echo intensity model indicators
    No.Material 1: PolyimideMaterial 2: SteelMaterial 3: WhiteMaterial 4: RedMaterial 5: BlueMaterial 6: Black
    13.09, 1.0, 80.52.81, 1.0, 1272.77, 1.0, 141.12.97, 1.0, 110.62.93, 1.0, 45.72.99, 1.0, 15.4
    22.91, 1.0, 168.32.75, 1.0, 145.52.63, 1.0, 153.62.81, 1.0, 158.52.75, 1.0, 70.52.75, 1.0, 38.1
    32.37, 1.0, 106.22.51, 1.0, 109.82.49, 1.0, 112.52.51, 1.0, 144.52.61, 1.0, 64.12.51, 1.0, 22.5
    42.77, 1.2, 119.32.73, 1.2, 69.32.57, 1.2, 119.82.79, 1.2, 76.22.77, 1.2, 22.12.83, 1.2, 7.8
    52.27, 1.2, 136.62.59, 1.2, 117.92.37, 1.2, 106.52.55, 1.2, 131.62.65, 1.2, 43.22.59, 1.2, 24.8
    62.19, 1.2, 91.12.37, 1.2, 94.92.31, 1.2, 84.32.39, 1.2, 118.32.39, 1.2, 62.32.29, 1.2, 15.5
    72.71, 1.4, 101.22.61, 1.4, 63.92.61, 1.4, 81.92.65, 1.4, 79.82.63, 1.4, 28.92.69, 1.4, 9.6
    82.65, 1.4, 141.72.51, 1.4, 91.22.45, 1.4, 118.92.51, 1.4, 125.32.45, 1.4, 57.62.43, 1.4, 22.1
    92.11, 1.4, 134.32.25, 1.4, 73.42.25, 1.4, 103.92.23, 1.4, 104.22.19, 1.4, 41.52.15, 1.4, 14.1
    102.49, 1.5, 135.32.38, 1.5, 54.62.62, 1.5, 34.32.49, 1.5, 70.12.57, 1.5, 27.92.53, 1.5, 9.9
    112.11, 1.5, 149.72.12, 1.5, 47.72.33, 1.5, 82.52.41, 1.5, 87.52.43, 1.5, 46.12.39, 1.5, 18.2
    122.05, 1.5, 112.22.04, 1.5, 35.282.19, 1.5, 73.82.07, 1.5, 55.62.13, 1.5, 36.52.11, 1.5, 13.7
    132.59, 1.6, 32.52.47, 1.6, 30.12.43, 1.6, 45.52.47, 1.6, 43.92.45, 1.6, 14.32.45, 1.6, 9.1
    142.47, 1.6, 72.12.29, 1.6, 55.62.31, 1.6, 62.72.35, 1.6, 67.12.27, 1.6, 28.82.25, 1.6, 18.8
    151.93, 1.6, 62.42.05, 1.6, 41.82.03, 1.6, 53.11.97, 1.6, 57.92.15, 1.6, 30.91.83, 1.6, 6.9
    162.27, 2.0, 14.62.13, 2.0, 10.12.07, 2.0, 12.62.03, 2.0, 15.392.05, 2.0, 6.92.15, 2.0, 2.2
    172.15, 2.0, 27.42.03, 2.0, 15.71.99, 2.0, 19.21.85, 2.0, 28.61.89, 2.0, 13.41.93, 2.0, 6.25
    181.37, 2.0, 9.31.87, 2.0, 26.11.87, 2.0, 26.731.65, 2.0, 21.871.75, 2.0, 15.51.53, 2.0, 5.4
    Table 3. Laser echo intensity data of unknown material surface
    No.Angle/(°)Distance/mmSSA-GRNNGRNN
    Predicted intensityEsquaredPredicted materialPredicted intensityEsquaredPredicted material
    13.09100079.1651.782178.4034.3974
    22.911000173.38525.8571172.84420.6481
    32.371000108.5265.41192.871177.6624
    42.771200120.2540.911129.98114.0621
    52.271200133.7987.8511135.9730.3931
    62.19120088.4716.912179.392137.0774
    72.711400102.1070.823172.516822.7724
    82.651400141.9430.0591155.155181.0371
    92.111400127.93440.5261126.84555.5771
    102.59160032.7720.074129.23310.6733
    112.47160072.5330.188180.71774.2531
    121.93160061.3221.162147.546220.6414
    132.272 00014.3090.085112.0236.6412
    142.152 00026.6900.504137.38499.681
    151.372 0009.8060.256110.8072.2714
    Table 4. Inversion results of SSA-GRNN model for material 1
    Ronghua LI, Xinchen ZHOU, Chuanxin WENG, Haopeng XUE, Jinlong WU, Chenyu LIN. Non-contact infrared laser physical property inversion method for target surface based on SSA-GRNN[J]. Infrared and Laser Engineering, 2024, 53(10): 20240217
    Download Citation