Author Affiliations
College of Equipment Management and Support, Engineering University of PAP, Xi’an, Shaanxi 710086, Chinashow less
Fig. 1. Structure of the CNN
Fig. 2. Flow chart of the SAR image target recognition based on updated classifier
Fig. 3. Target images in the MSTAR data set. (a) BMP2; (b) BTR70; (c) T72; (d) T62; (e) BRDM2; (f) BTR60; (g) ZSU23/4; (h) D7; (i) ZIL131; (j) 2S1
Fig. 4. Classification results under standard operating conditions
Fig. 5. Classification results of different methods under noise interference
Fig. 6. Comparison of results under small traning samples
Class | Training (17°) | Test (15°) |
---|
Configuration | Number of samples | Configuration | Number of samples |
---|
BMP2 | 9563 | 233 | 95639566c21 | 195196196 | BTR70 | -- | 233 | -- | 196 | T72 | 132 | 232 | 132812S7 | 196195191 | T62 | -- | 299 | -- | 273 | BRDM2 | -- | 298 | -- | 274 | BTR60 | -- | 256 | -- | 195 | ZSU23/4 | -- | 299 | -- | 274 | D7 | -- | 299 | -- | 274 | ZIL131 | -- | 299 | -- | 274 | 2S1 | -- | 299 | -- | 274 |
|
Table 1. Test scenarios for standard operating conditions
Method | Average recognition rate |
---|
Ours | 99.08 | SRC | 96.42 | A-ConvNet | 98.53 | Aug-CNN | 98.91 | SVM+SRC | 97.48 |
|
Table 2. Classification results of different methods under standard operating conditions unit: %
Threshold | 1.1 | 1.2 | 1.3 | 1.4 | 1.5 | 1.6 | 1.7 |
---|
Average recognition rate | 98.62 | 98.86 | 99.02 | 99.08 | 99.03 | 98.79 | 98.54 |
|
Table 3. Classification results of our method under different decision reliability thresholds unit: %
Class | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
---|
Decision value 1 | 0.28 | 0.05 | 0.32 | 0.04 | 0.10 | 0.04 | 0.01 | 0.06 | 0.05 | 0.05 | Decision value 2 | 0.36 | 0.07 | 0.18 | 0.06 | 0.08 | 0.07 | 0.02 | 0.05 | 0.07 | 0.04 | Fused decision value | 0.32 | 0.06 | 0.25 | 0.05 | 0.05 | 0.06 | 0.02 | 0.06 | 0.06 | 0.05 |
|
Table 4. Decision variable distribution of BMP2 test sample
Test subset | Part 1 | Part 2 | Part 3 | Part 4 | All |
---|
Average recognition rate | 69.32 | 75.45 | 85.64 | 89.12 | 79.86 |
|
Table 5. Average recognition rates of different test subsets unit: %
Class | Training | Test |
---|
Depression | Number of samples | Depression | Number of samples |
---|
2S1 | 17° | 299 | 30°45° | 288303 | BDRM2 | 17° | 298 | 30°45° | 287303 | ZSU23/4 | 17° | 299 | 30°45° | 288303 |
|
Table 6. Test conditions for pitch angle difference
Method | Depression |
---|
30° | 45° |
---|
Ours | 96.12 | 72.74 | SRC | 92.17 | 64.38 | A-ConvNet | 94.12 | 65.93 | Aug-CNN | 95.38 | 69.32 | SVM+SRC | 94.16 | 66.08 |
|
Table 7. Test results under different pitch angles unit: %