• Spacecraft Recovery & Remote Sensing
  • Vol. 45, Issue 5, 147 (2024)
Wenzhuo WANG1, Chenglong MA2, Guanlin WANG1, Yiming ZHANG1..., Fangxiong TAN2,*, Xu HAN3 and Lei WU3|Show fewer author(s)
Author Affiliations
  • 1State Grid Gansu Electric Power Company, Lanzhou 730000, China
  • 2State Grid Gansu Electric Power Company Jiuquan Power Supply Company, Jiuquan 735000, China
  • 3Beijing Deep Blue Space Remote Sensing Technology Co., Ltd., Beijing 100020, China
  • show less
    DOI: 10.3969/j.issn.1009-8518.2024.05.014 Cite this Article
    Wenzhuo WANG, Chenglong MA, Guanlin WANG, Yiming ZHANG, Fangxiong TAN, Xu HAN, Lei WU. Real-Time Fire Detection by Cascading Traditional Approaches with Deep Learning[J]. Spacecraft Recovery & Remote Sensing, 2024, 45(5): 147 Copy Citation Text show less
    [in Chinese]
    Fig. 1. [in Chinese]
    [in Chinese]
    Fig. 2. [in Chinese]
    [in Chinese]
    Fig. 3. [in Chinese]
    [in Chinese]
    Fig. 4. [in Chinese]
    类别波段序号中心波长/μm空间分辨率/km
    可见光R10.471.0
    R20.511.0
    R30.640.5
    近红外R40.861.0
    R51.62.0
    R62.32.0
    红外T73.92.0
    T86.22.0
    T96.92.0
    T107.32.0
    T118.62.0
    T129.62.0
    T1310.42.0
    T1411.22.0
    T1512.42.0
    T1613.32.0
    Table 1. Himawari-8 band introduction
    分组类型输入特征
    1原始光谱信息R1R2R3R4R5R6T7T8
    T9T10T11T12T13T14T15T16
    2亮温差值T7-T8T7-T9T7-T10T7-T11T7-T12T7-T13T7-T14T7-T15T7-T16
    T8-T9T8-T10T8-T11T8-T12T8-T13T8-T14T8-T15T8-T16
    T9-T10T9-T11T9-T12T9-T13T9-T14T9-T15T9-T16
    T10-T11T10-T12T10-T13T10-T14T10-T15T10-T16
    T11-T12T11-T13T11-T14T11-T15T11-T16
    T12-T13T12-T14T12-T15T12-T16
    T13-T14T13-T15T13-T16
    T14-T15T14-T16
    T15-T16
    亮温比值T7/T8T7/T9T7/T10T7/T11T7/T12T7/T13T7/T14T7/T15T7/T16
    T8/T9T8/T10T8/T11T8/T12T8/T13T8/T14T8/T15T8/T16
    T9/T10T9/T11T9/T12T9/T13T9/T14T9/T15T9/T16
    T10/T11T10/T12T10/T13T10/T14T10/T15T10/T16
    T11/T12T11/T13T11/T14T11/T15T11/T16
    T12/T13T12/T14T12/T15T12/T16
    T13/T14T13/T15T13/T16
    T14/T15T14/T16
    T15/T16
    3空间上下文信息MEAN_T7MEAN_T14MEAN_BT7
    STD_T7STD_T14STD_BT7
    4地理差异DEMSlopeAspectLanduseNDVILonLat
    时间差异DOYHour_minute
    观测角度SAZSAAS0ZSOA
    Table 2. Model input features
    类别预测为火点预测为非火点
    真实火点TPFN
    真实非火点FPTN
    Table 3. Confusion matrix
    模型$ \mathit{P} $$ \mathit{M} $$ \mathit{F} $
    Model1-CNN0.720.270.73
    Model2-CNN0.770.180.79
    Model3-CNN0.830.140.84
    PMCNN0.810.140.83
    MCNN0.880.110.88
    Table 4. Model prediction results
    类别识别火点总数/个正确数/个漏报数/个$ P $$ M $$ F $
    MCNN10388180.850.170.84
    WLF82240.23
    Table 5. Comparison of fire point recognition accuracy
    Wenzhuo WANG, Chenglong MA, Guanlin WANG, Yiming ZHANG, Fangxiong TAN, Xu HAN, Lei WU. Real-Time Fire Detection by Cascading Traditional Approaches with Deep Learning[J]. Spacecraft Recovery & Remote Sensing, 2024, 45(5): 147
    Download Citation