• Acta Optica Sinica
  • Vol. 42, Issue 2, 0210002 (2022)
Shijie Deng1, Haiyan Wang1、*, An Xu1, Chunqing Gao2, and Junbing Li1
Author Affiliations
  • 1College of Aeronautics and Astronautics Engineering, Air Force Engineering University, Xian, Shaanxi 710038, China
  • 2Unit 94582 of the Chinese Peoples Liberation Army, Shangqiu, Henan 476000, China
  • show less
    DOI: 10.3788/AOS202242.0210002 Cite this Article Set citation alerts
    Shijie Deng, Haiyan Wang, An Xu, Chunqing Gao, Junbing Li. Target Detection Method Based on Antigrowth[J]. Acta Optica Sinica, 2022, 42(2): 0210002 Copy Citation Text show less
    Flow of proposed algorithm
    Fig. 1. Flow of proposed algorithm
    Experimental data. (a) Original image; (b) objectives; (c) synthesis of data
    Fig. 2. Experimental data. (a) Original image; (b) objectives; (c) synthesis of data
    Detection effects of different algorithms at different thresholds. (a)--(c) SAM algorithm; (d) proposed algorithm
    Fig. 3. Detection effects of different algorithms at different thresholds. (a)--(c) SAM algorithm; (d) proposed algorithm
    Time complexity of proposed algorithm under different initial number of points
    Fig. 4. Time complexity of proposed algorithm under different initial number of points
    Target growth process. (a) 1st layer; (b) 2nd layer; (c) 3rd layer; (d) 4th layer; (e) dentification result
    Fig. 5. Target growth process. (a) 1st layer; (b) 2nd layer; (c) 3rd layer; (d) 4th layer; (e) dentification result
    Root node optimization process. (a) Short chain growth result; (b) adjacent frequency
    Fig. 6. Root node optimization process. (a) Short chain growth result; (b) adjacent frequency
    Background growth process. (a) 1st layer; (b) 2nd layer; (c) 3rd layer; (d) 4th layer; (e) against result
    Fig. 7. Background growth process. (a) 1st layer; (b) 2nd layer; (c) 3rd layer; (d) 4th layer; (e) against result
    Position of missing and overlapping pixels and final recognition effect. (a) Comprehensive display effect; (b) missing pixels; (c) overlapping pixels; (d) determination effect
    Fig. 8. Position of missing and overlapping pixels and final recognition effect. (a) Comprehensive display effect; (b) missing pixels; (c) overlapping pixels; (d) determination effect
    Detection effects of different algorithms. (a) Original images; (b) truth ground; (c) CEM algorithm; (d) ACE algorithm; (e) WCM-OSP algorithm; (f) DERSG algorithm; (g) AG algorithm
    Fig. 9. Detection effects of different algorithms. (a) Original images; (b) truth ground; (c) CEM algorithm; (d) ACE algorithm; (e) WCM-OSP algorithm; (f) DERSG algorithm; (g) AG algorithm
    ROCs of different algorithms under different datasets. (a) Data 1; (b) data 2; (c) data 3; (d) data 4
    Fig. 10. ROCs of different algorithms under different datasets. (a) Data 1; (b) data 2; (c) data 3; (d) data 4
    AlgorithmThresholdPd /%Pf /%
    C=0.052014.030
    SAMC=0.070036.840.08
    C=0.078445.332.05
    Proposed algorithmC1=0.0520,C2=0.070045.330.08
    Table 1. Detection probability and false alarm probability of different algorithms
    DataParameterCEMACEWCM-OSPAGDESRG
    1Pd /%90.00
    Pf /%0.110.2513.480.0734.00
    2Pd /%90.00
    Pf /%0.831.1057.320.7683.77
    3Pd /%90.00
    Pf /%1.186.321.730.090.34
    4Pd /%90.00
    Pf /%0.260.284.560.2015.34
    Table 2. Detection probability and false alarm rate of different algorithms when detection probability is 90%
    ParameterCEMACEWCM-OSPAGDESRG
    Average time /s6.4831.376.6743.5932.28
    Table 3. Average running time of different algorithms when detection probability is 90%
    Shijie Deng, Haiyan Wang, An Xu, Chunqing Gao, Junbing Li. Target Detection Method Based on Antigrowth[J]. Acta Optica Sinica, 2022, 42(2): 0210002
    Download Citation