Anti-interference recognition method of aerial infrared targets based on a spatio-temporal correlation inference network

Liang Zhang; Xiaoqian Tian; Shaoyi Li; Xi Yang

doi:10.3788/IRLA20210614

Journals >Infrared and Laser Engineering >Volume 51 >Issue 7 >Page 20210614 > Article

Infrared and Laser Engineering
Vol. 51, Issue 7, 20210614 (2022)

Anti-interference recognition method of aerial infrared targets based on a spatio-temporal correlation inference network

Liang Zhang^1,2, Xiaoqian Tian³, Shaoyi Li^3,*, and Xi Yang³

Author Affiliations

¹China Airborne Missile Academy, Luoyang 471009, China

²Aviation Key Laboratory of Science and Technology on Airborne Guided Weapon, Luoyang 471009, China

³School of Astronautics, Northwestern Polytechnical University, Xi’an 710072, China

show less

DOI: 10.3788/IRLA20210614 Cite this Article

Liang Zhang, Xiaoqian Tian, Shaoyi Li, Xi Yang. Anti-interference recognition method of aerial infrared targets based on a spatio-temporal correlation inference network[J]. Infrared and Laser Engineering, 2022, 51(7): 20210614 Copy Citation Text

show less

Fig. 1. Representation diagram of space-time correlation inference network

Download full size | View in the Article

Fig. 2. Framework diagram of anti-interference recognition algorithm based on space-time correlation inference network

Download full size | View in the Article

Fig. 3. Schematic diagram of target and interference sample labeling

Download full size | View in the Article

Fig. 4. Initial network structure diagram

Download full size | View in the Article

Fig. 5. Network structure diagram based on space-time correlation inference

Download full size | View in the Article

Fig. 6. Test chart of target recognition algorithm based on space-time correlation inference network

Download full size | View in the Article

Total decoys	Projection group number	Group interval/s	Number of decoys per group	Decoys interval/s	Maneuver
24	24	1	1	0.1	Without maneuver, turn left, jump
24	12	1	2	0.1	Without maneuver, turn left, jump
24	6	1	4	0.1	Without maneuver, turn left, jump

Table 1. Interference projection strategy

View in the Article

C=1
	Y1	Y2	Y3		Y25
Z1	0	0	0	…	0
Z2	0.220 0	0.002 5	0	…	0
Z3	0.778 8	0.849 7	0.156 3	…	0
Z4	0.001 2	0.147 8	0.673 7	…	0
Z5	0	6.749 3e-05	0.168 2	…	0
…	…	…	…	…	…
Z29	0	0	0	…	0

C=0
	Y1	Y2	Y3		Y25
Z1	0.475 9	0.039 0	0	…	0
Z2	0.392 0	0.479 5	0.047 3	…	0
Z3	0.129 3	0.328 3	0.401 9	…	0
Z4	0.002 9	0.105 2	0.384 1	…	0
Z5	0	0.048 0	0.070 3	…	0
…	…	…	…	…	…
Z29	0	0	0	…	0

Table 2. Time slice conditional probability table of area feature node

View in the Article

C=1
注：片内概率表面积的父节点是周长；片间转移概率表面积的父节点是当前时刻的周长以及上一时刻的面积。面积特征节点的转移概率表是三维概率表，表3选取了当前时刻面积的第五个特征区间对应的父节点的转移概率表。
	Y1	Y2	Y3		Y25
Z1	0	0	0	…	0
Z2	0	0	0.162 4	…	0
Z3	0	5.419 84-06	0.008 4	…	0
Z4	0	3.210 0e-05	0.041 3	…	0
Z5	0	0.975 6	0.945 9	…	0
…	…	…	…	…	…
Z29	0	0	0	…	0

C=0
	Y1	Y2	Y3		Y25
Z1	0	0.006 7	0.007 3	…	0
Z2	0	0.004 3	0.042 8	…	0
Z3	0	4.022 5e-04	0.056 1	…	0
Z4	0	4.161 5e-04	0.027 1	…	0
Z5	0	0.027 3	0.051 8	…	0
…	…	…	…	…	…
Z29	0	0	0	…	0

Table 3. Transition probability table of area feature node

View in the Article

Launch distance/m	Relative azimuth	Type of maneuver	Launch conditions			TAN algorithm accuracy	Algorithm accuracy of this paper
Launch distance/m	Relative azimuth	Type of maneuver	Firing interval	Number of decoys launched	Number of launch groups	TAN algorithm accuracy	Algorithm accuracy of this paper
7000	10°	Turn left	0.4	2	24	90.73	93.29
7000	10°	Jump	0.4	2	24	87.22	94.79
7000	10°	Turn left	0.4	4	12	83.06	92.91
7000	10°	Jump	0.4	4	12	88.18	93.93
7000	10°	Turn left	0.7	2	24	93.95	92.54
7000	10°	Jump	0.7	2	24	90.54	94.41
7000	10°	Turn left	0.7	4	12	86.72	94.72
7000	10°	Jump	0.7	4	12	91.54	94.83
7000	40°	Turn left	0.4	2	24	89.22	93.13
7000	40°	Jump	0.4	2	24	93.23	94.22
7000	40°	Turn left	0.4	4	12	85.22	94.56
7000	40°	Jump	0.4	4	12	93.48	95.05
7000	40°	Turn left	0.7	2	24	91.92	93.87
7000	40°	Jump	0.7	2	24	95.88	96.87
7000	40°	Turn left	0.7	4	12	86.82	91.07
7000	40°	Jump	0.7	4	12	96.13	95.16
7000	100°	Turn left	0.4	2	24	85.51	92.32
7000	100°	Jump	0.4	2	24	95.77	95.93
7000	100°	Turn left	0.4	4	12	94.65	95.33
7000	100°	Jump	0.4	4	12	95.32	96.55
7000	100°	Turn left	0.7	2	24	85.92	92.54
7000	100°	Jump	0.7	2	24	98.14	94.60
7000	100°	Turn left	0.7	4	12	97.07	96.63
7000	100°	Jump	0.7	4	12	97.75	94.81
7000	160°	Turn left	0.4	2	24	92.50	95.15
7000	160°	Jump	0.4	2	24	91.25	91.91
7000	160°	Turn left	0.4	4	12	94.75	95.06
7000	160°	Jump	0.4	4	12	92.00	93.72
7000	160°	Turn left	0.7	2	24	95.04	96.01
7000	160°	Jump	0.7	2	24	93.75	94.14
7000	160°	Turn left	0.7	4	12	97.33	97.42
7000	160°	Jump	0.7	4	12	94.57	95.57

Table 4. Airborne infrared target recognition algorithm test results of two algorithms

Download Citation

Save the article for my favorites

Paper Information

微信扫一扫：分享

微信扫一扫：分享