Infrared Dim-Small Target Detection Based on Robust Principal Component Analysis and Multi-Point Constant False Alarm

Mingyang Ma; Dejiang Wang; He Sun; Tao Zhang

doi:10.3788/AOS201939.0810001

Acta Optica Sinica, Volume. 39, Issue 8, 0810001(2019)

Infrared Dim-Small Target Detection Based on Robust Principal Component Analysis and Multi-Point Constant False Alarm

Mingyang Ma^1,2, Dejiang Wang^1、*, He Sun¹, and Tao Zhang¹

¹ Key Laboratory of Aviation Optical Imaging and Measurement, Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun, Jilin 130033, China

² School of Optoelectronics, University of Chinese Academy of Sciences, Beijing 100049, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(14)

Fig. 1. Energy diffusion of small target. (a) Transverse diffusion; (b) non-diffusion; (c) peripheral diffusion; (d)-(f) three-dimensional distributions of Figs. 1(a)-(c), respectively

Download full size

Fig. 2. SNR of small target in successive 100 frames

Download full size

Fig. 3. Flow chart of proposed method

Download full size

Fig. 4. Constant false alarm detection windows of different pixels. (a) Single pixel; (b) multi-pixel

Download full size

Fig. 5. Typical energy distribution of target when M=3

Download full size

Fig. 6. Effect of M value on probability of detection

Download full size

Fig. 7. Effect of SNR on probability of detection at different M values. (a) M=3; (b) M=4

Download full size

Fig. 8. Results of proposed algorithm. (a)-(d) Original images; (e)-(h) results of improved threshold segmentation algorithm based on RPCA; (i)-(l) results of multi-point constant false alarm detection

Download full size

Fig. 9. Processing results of different algorithms. (a)(b) Original images; (c)(d) improved algorithm based on RPCA; (e)(f) top hat transformation algorithm; (g)(h) BM3D algorithm

Download full size

Fig. 10. Results of different constant false alarm detection algorithms. (a)(b) Original images; (c)(d) MCFAR algorithm; (e)(f) CFAR algorithm; (g)(h) ACFAR algorithm

Download full size

Table 1. Related parameters of targets in Fig. 8
View table
Table 1. Related parameters of targets in Fig. 8
ParameterR_sn Fig. 8(a) Fig. 8(b) Fig. 8(c) Fig. 8(d)
1.827 2.78 2.227 2.347 3.72 3.88
Location (269,143) (316,182) (242,91) (312,127) (79,93) (105,219)

Table 2. Related parameters of targets in Fig. 9
View table
Table 2. Related parameters of targets in Fig. 9
ParameterR_sn Fig. 9(a) Fig. 9(b)
2.23 2.35 2.57 2.23
Location (242,91) (312,127) (239,380) (314,420)

Table 3. Number of candidate points for segmentation algorithms
View table
Table 3. Number of candidate points for segmentation algorithms
Frame Improved RPCA Top hat algorithm BM3D algorithm
Fig. 9(a) 34 5001 1249
Fig. 9(b) 22 5353 1143

Table 4. Results of different constant false alarm detection algorithms
View table
Table 4. Results of different constant false alarm detection algorithms
Method R_DR /% R_FAR /%
Improved RPCA+CFAR 80.0 93.9
Improved RPCA+ACFAR 67.6 84.9
Improved RPCA+MCFAR 95.6 37.8

Tools

Get Citation

Copy Citation Text

Mingyang Ma, Dejiang Wang, He Sun, Tao Zhang. Infrared Dim-Small Target Detection Based on Robust Principal Component Analysis and Multi-Point Constant False Alarm[J]. Acta Optica Sinica, 2019, 39(8): 0810001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites