• Laser & Optoelectronics Progress
  • Vol. 59, Issue 24, 2410010 (2022)
Yanfang Yan1, Qing Wang1,*, Qihong Zeng2, Yanlin Shao1..., Wei Wei1 and Changmin Zhang1|Show fewer author(s)
Author Affiliations
  • 1School of Geosciences, Yangtze University, Wuhan 430100, Hubei, China
  • 2China Institute of Petroleum Exploration and Development, Beijing 100083, China
  • show less
    DOI: 10.3788/LOP202259.2410010 Cite this Article Set citation alerts
    Yanfang Yan, Qing Wang, Qihong Zeng, Yanlin Shao, Wei Wei, Changmin Zhang. Intelligent Identification of Clastic Rock Outcrops from Multimodal UAV Images[J]. Laser & Optoelectronics Progress, 2022, 59(24): 2410010 Copy Citation Text show less
    RGB-D semantic segmentation for clastic rocks
    Fig. 1. RGB-D semantic segmentation for clastic rocks
    SE-DeepLabv3+ network structure
    Fig. 2. SE-DeepLabv3+ network structure
    SE module
    Fig. 3. SE module
    Overview of the study area
    Fig. 4. Overview of the study area
    Overall process of clastic rock lithology identification
    Fig. 5. Overall process of clastic rock lithology identification
    Comparison of recognition results of different methods
    Fig. 6. Comparison of recognition results of different methods
    Comparison of recognition results with other semantic segmentation methods
    Fig. 7. Comparison of recognition results with other semantic segmentation methods
    Recognition result for some outcrops in Qingshuihe-Kalaza Formation
    Fig. 8. Recognition result for some outcrops in Qingshuihe-Kalaza Formation
    MethodOA /%mIOU /%
    Maximum likelihood method63.1625.04
    Mahalanobis distance method59.0123.94
    Minimum distance method52.9016.34
    K-means28.4013.11
    Isodata28.4013.11
    SE-DeepLabv3+91.0536.92
    Table 1. Comparison of recognition accuracy of different methods
    MethodLithologyPA /%OA /%mIOU /%Prediction time /s
    FCNConglomerate61.7772.5223.521748.84
    Sandstone0.02
    Mudstone93.59
    U-NetConglomerate74.7681.1127.67120.88
    Sandstone9.17
    Mudstone36.34
    DeepLabv3+Conglomerate94.0574.6926.52130.85
    Sandstone6.21
    Mudstone65.73
    SE-DeepLabv3+Conglomerate74.3391.0536.92146.40
    Sandstone50.03
    Mudstone83.13
    Table 2. Comparison of recognition accuracy of different semantic segmentation methods
    MethodLithologyPA /%OA /%mIOU /%Prediction time /s
    SE-DeepLabv3+Conglomerate50.0280.6034.41798.99
    Sandstone40.76
    Mudstone80.49
    Table 3. Recognition accuracy for some outcrops in Qingshuihe-Kalaza Formation
    Yanfang Yan, Qing Wang, Qihong Zeng, Yanlin Shao, Wei Wei, Changmin Zhang. Intelligent Identification of Clastic Rock Outcrops from Multimodal UAV Images[J]. Laser & Optoelectronics Progress, 2022, 59(24): 2410010
    Download Citation