Dim Target Detection in Airborne Infrared Images Based on Visual Feature Fusion

Fig. 2. Distribution maps of intensity and gradient of target and its neighborhood. (a) Original image; (b) 3D intensity distribution; (c) gradient map

Download full size

Fig. 3. Cloud and its gradient distribution. (a) Original image; (b) gradient distribution of cloud edge region

Download full size

Fig. 4. Gradient operators in four directions

Download full size

Fig. 5. Detection results. (a) Original image; (b) 3D intensity distribution of original image; (c) effect image; (d) 3D intensity distribution of effect image

Download full size

Fig. 6. Detection results of different methods. (a) Original image; (b) Top-Hat method; (c) LRES method; (d) NRAM method; (e) LIG method; (f) proposed method

Download full size

Table 1. Detection performance of different methods
View table
Table 1. Detection performance of different methods
Method Top-Hat LRES NRAM LIG Proposed method
Detection rate 0.82 0.86 0.89 0.85 0.91
False detection rate 0.15 0.16 — 0.09 0.06

Table 2. SCR of different methods
View table
Table 2. SCR of different methods
Sequency Top-Hat LRES NRAM LIG Proposed method
1 3.5400 6.2757 6.5206 6.0686 7.7595
2 4.3554 4.8404 4.9644 5.5557 6.0472
3 2.1978 2.7258 — 2.6150 3.1563
4 3.8955 4.7288 — 4.8650 6.3940

Table 3. BSF of different methods
View table
Table 3. BSF of different methods
Sequency Top-Hat LRES NRAM LIG Proposed method
1 686.0928 3414.1000 3638.8 2094.4000 3741.1
2 398.0309 922.5906 1176.9 1020.7000 1156.7
3 502.2085 3307.4000 — 3321.7000 4084.9
4 483.2480 714.7317 — 537.9541 4215.1

Tools

Get Citation

Copy Citation Text

Guoqing Qiu, Haijing Yang, Yantao Wang, Yating Wei, Pan Luo. Dim Target Detection in Airborne Infrared Images Based on Visual Feature Fusion[J]. Laser & Optoelectronics Progress, 2020, 57(18): 181004

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites