Infrared Small Target Detection Using Gradient Differential Anisotropic Gaussian Filtering

Benchen Yang; Wanni Song; Haibo Jin; Simeng Li

doi:10.3788/LOP222723

Journals >Laser & Optoelectronics Progress >Volume 60 >Issue 16 >Page 1612003 > Article

Laser & Optoelectronics Progress
Vol. 60, Issue 16, 1612003 (2023)

Infrared Small Target Detection Using Gradient Differential Anisotropic Gaussian Filtering

Benchen Yang, Wanni Song^*, Haibo Jin, and Simeng Li

Author Affiliations

School of Software, Liaoning Technical University, Xingcheng125100, Huludao, China

show less

DOI: 10.3788/LOP222723 Cite this Article Set citation alerts

Benchen Yang, Wanni Song, Haibo Jin, Simeng Li. Infrared Small Target Detection Using Gradient Differential Anisotropic Gaussian Filtering[J]. Laser & Optoelectronics Progress, 2023, 60(16): 1612003 Copy Citation Text

show less

Fig. 1. Schematic diagram of the proposed algorithm

Download full size

Fig. 2. Gaussian template

Download full size

Fig. 3. Sliding window structure

Download full size

Fig. 4. ROC curves of different values of k

Download full size

Fig. 5. Experiment comparison figure 1

Download full size

Fig. 6. Experiment comparison figure 2

Download full size

Fig. 7. Experiment comparison figure 3

Download full size

Fig. 8. Experiment comparison figure 4

Download full size

Fig. 9. Experiment comparison figure 5

Download full size

Fig. 10. Experiment comparison figure 6

Download full size

Fig. 11. ROC curves of different algorithms. (a) Complex scenes; (b) strong clutter scenes ; (c) highlight scenes

Download full size

No.	Resolution	Number	$R_{S C R}$	Target size	Background type	Scene description
1	261 $\times$ 201	2	3.002	3×3，4×4	Air	Multi-target+complex clouds+highlighted
2	328 $\times$ 220	2	1.710	4×4，5×5	Sea surface	Multi-target+strong clutter+highlighted
3	330 $\times$ 204	1	4.085	5×5	Air	Dark target+high noise+highlight edges
4	320 $\times$ 255	1	2.425	3×3	Air	Dark target+complex cloud+bright edges
5	269 $\times$ 235	1	3.803	5×5	Sea and air	Contour target+strong interference
6	286×188	2	3.977	2×2，3×3	Air	Multi-target+edge target

Table 1. infra-red image details

No.	Index	MLCM	LIG	TTLCM	WSLCM	PSTNN	FKRW	Proposed algorithm
1	$G_{S C R G}$	8.137	61.273	12.571	99.022	100.675	99.874	135.529
1	$F_{B S F}$	1.326	19.189	2.915	10.946	10.762	8.385	38.388
2	$G_{S C R G}$	10.806	13.245	47.743	25.034	27.582	68.739	87.558
2	$F_{B S F}$	0.523	0.466	2.673	0.996	2.875	1.487	18.129
3	$G_{S C R G}$	7.369	1.092	22.301	22.365			197.564
3	$F_{B S F}$	0.092	2.089	0.226	0.179			31.316
4	$G_{S C R G}$	4.995	2.211	8.720	24.757			69.5651
4	$F_{B S F}$	0.653	21.569	0.494	1.286			18.763
5	$G_{S C R G}$	2.922	4.837	40.099	0.046	10.463	37.892	109.515
5	$F_{B S F}$	0.437	0.767	4.116	9.206	3.291	2.351	16.252
6	$G_{S C R G}$	3.426	5.973	32.367	1.724	9.624	25.576	79.263
6	$F_{B S F}$	2.387	4.658	7.384	12.856	5.928	10.592	31.374

Table 2.

G_{S C R G}

and

F_{B S F}

for different algorithms

Algorithm	Time
MLCM	0.2175
LIG	0.6075
TTLCM	0.3949
WSLCM	0.9565
PSTNN	0.6652
FKRW	0.1063
Proposed algorithm	0.5911

Table 3. Real-time comparison of different algorithm

Benchen Yang, Wanni Song, Haibo Jin, Simeng Li. Infrared Small Target Detection Using Gradient Differential Anisotropic Gaussian Filtering[J]. Laser & Optoelectronics Progress, 2023, 60(16): 1612003

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information

微信扫一扫：分享

微信扫一扫：分享