Author Affiliations
11. The 59th Research Institute of China Ordnance Industry, Chongqing 400039, China22. Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, Chinashow less
Fig. 1. Test scenarios
(a): Grassland background and vegetation camouflage target; (b): Soil background and vegetation camouflage target; (c): Soil background and vegetation and cement road camouflage targets; (d): Grassland/cement road/soil background and its corresponding camouflage targets
Fig. 2. Detection results of vegetation camouflage targets under grassland background using different algorithms
(a): MtACE; (b): MtAMF; (c): MtCEM; (d): SumACE; (e): SumAMF; (f): SumCEM; (g): WtaACE; (h): WtaAMF; (i): WtaCEM
Fig. 3. Detection results of vegetation camouflage targets under soil background using different algorithms
(a): MtACE; (b): MtAMF; (c): MtCEM; (d): SumACE; (e): SumAMF; (f): SumCEM; (g): WtaACE; (h): WtaAMF; (i): WtaCEM
Fig. 4. Detection results of vegetation and soil camouflage targets under soil background using different algorithms
(a): MtACE; (b): MtAMF; (c): MtCEM; (d): SumACE; (e): SumAMF; (f): SumCEM; (g): WtaACE; (h): WtaAMF; (i): WtaCEM
Fig. 5. Detection results of vegetation/cement road/soil camouflage targets under grassland/cement road/soil background using different algorithms
(a): MtACE; (b): MtAMF; (c): MtCEM; (d): SumACE; (e): SumAMF; (f): SumCEM; (g): WtaACE; (h): WtaAMF; (i): WtaCEM
Fig. 6. Detection results of four scenarios by nine multiple features detection algorithms
Fig. 7. Detection results of three SNR scenarios by nine multiple features detection algorithms
Number | Test Schemes | Similarity between targets | Similarity between target and background |
---|
1 | Detection of vegetation camouflage targets under grassland background | large | large | 2 | Detection of vegetation camouflage targets under soil background | large | small | 3 | Detection of vegetation and road camouflage targets under soil background | small | small | 4 | Detection of vegetation, soil, road camouflage targets under grassland, soil, road background | small | large |
|
Table 1. Test scheme
Number | Algorithms | FD | Accuracy |
---|
1 | MtACE | 0.315 3 | 1.0 | 2 | MtAMF | 0.394 9 | 0.921 8 | 3 | MtCEM | 0.736 1 | 0.865 5 | 4 | SumACE | 0.391 5 | 0.931 1 | 5 | SumAMF | 0.391 5 | 0.931 1 | 6 | SumCEM | 0.858 8 | 0.864 2 | 7 | WtaACE | 0.666 8 | 0.683 6 | 8 | WtaAMF | 0.666 8 | 0.683 6 | 9 | WtaCEM | 0.937 4 | 0.642 1 |
|
Table 2. FD and accuracy parameter of test 1
Number | Algorithms | FD | Accuracy |
---|
1 | MtACE | 0.318 6 | 1.0 | 2 | MtAMF | 0.372 3 | 0.944 2 | 3 | MtCEM | 0.878 8 | 0.689 0 | 4 | SumACE | 0.366 6 | 0.956 4 | 5 | SumAMF | 0.366 6 | 0.956 4 | 6 | SumCEM | 0.861 1 | 0.864 0 | 7 | WtaACE | 0.744 5 | 0.753 5 | 8 | WtaAMF | 0.744 5 | 0.753 5 | 9 | WtaCEM | 0.964 3 | 0.593 5 |
|
Table 3. FD and accuracy parameter of test 2
Number | Algorithms | FD | Accuracy |
---|
1 | MtACE | 0.353 6 | 1.0 | 2 | MtAMF | 0.382 1 | 0.905 8 | 3 | MtCEM | 0.878 1 | 1.726 9 | 4 | SumACE | 0.430 2 | 0.940 2 | 5 | SumAMF | 0.430 2 | 0.940 2 | 6 | SumCEM | 0.914 7 | 0.805 5 | 7 | WtaACE | 0.691 6 | 0.712 5 | 8 | WtaAMF | 0.691 6 | 0.712 5 | 9 | WtaCEM | 0.986 8 | 0.457 3 |
|
Table 4. FD and accuracy parameter of test 3
Number | Algorithms | FD | Accuracy |
---|
1 | MtACE | 0.292 4 | 1.0 | 2 | MtAMF | 0.251 6 | 0.920 4 | 3 | MtCEM | 0.956 4 | 0.640 1 | 4 | SumACE | 0.275 0 | 0.956 2 | 5 | SumAMF | 0.275 0 | 0.956 2 | 6 | SumCEM | 0.749 8 | 0.936 3 | 7 | WtaACE | 0.616 9 | 0.632 5 | 8 | WtaAMF | 0.616 9 | 0.632 5 | 9 | WtaCEM | 0.904 3 | 0.796 8 |
|
Table 5. FD and accuracy parameter of test 4
Number | Algorithms | SNR=200 | SNR=200 | SNR=400 | SNR=400 | SNR=800 | SNR=800 |
---|
FD | Accuracy | FD | Accuracy | FD | Accuracy |
---|
1 | MtACE | 0.381 2 | 0.754 3 | 0.305 1 | 0.821 1 | 0.254 3 | 0.873 8 | 2 | MtAMF | 0.394 5 | 0.860 1 | 0.306 4 | 0.886 6 | 0.261 7 | 0.922 2 | 3 | MtCEM | 0.773 3 | 0.821 5 | 0.748 2 | 0.839 0 | 0.802 7 | 0.874 3 | 4 | SumACE | 0.449 6 | 0.842 4 | 0.361 2 | 0.873 0 | 0.297 9 | 0.917 9 | 5 | SumAMF | 0.449 6 | 0.842 4 | 0.361 2 | 0.873 0 | 0.297 9 | 0.917 9 | 6 | SumCEM | 0.841 8 | 0.791 0 | 0.843 6 | 0.813 3 | 0.857 2 | 0.857 7 | 7 | WtaACE | 0.691 5 | 0.701 1 | 0.695 9 | 0.705 8 | 0.699 5 | 0.699 8 | 8 | WtaAMF | 0.691 5 | 0.701 1 | 0.695 9 | 0.705 8 | 0.699 5 | 0.699 8 | 9 | WtaCEM | 0.956 6 | 0.574 1 | 0.962 9 | 0.552 7 | 0.961 4 | 0.562 4 |
|
Table 6. Detection results under different signal-to-noise ratios
Algorithms | Tests of spectrum | Tests of noise |
---|
Mean | Variance | Mean | Variance |
---|
MtACE | 0.320 0 | 0.021 9 | 0.313 5 | 0.052 1 | MtAMF | 0.350 2 | 0.057 5 | 0.320 9 | 0.055 2 | MtCEM | 0.862 4 | 0.079 5 | 0.774 7 | 0.022 3 | SumACE | 0.365 8 | 0.057 1 | 0.369 6 | 0.062 2 | SumAMF | 0.365 8 | 0.057 1 | 0.369 6 | 0.062 2 | SumCEM | 0.846 1 | 0.059 9 | 0.847 5 | 0.006 9 | WtaACE | 0.680 0 | 0.046 0 | 0.695 6 | 0.003 3 | WtaAMF | 0.680 0 | 0.046 0 | 0.695 6 | 0.003 3 | WtaCEM | 0.948 2 | 0.030 8 | 0.960 3 | 0.002 7 |
|
Table 7. Statistics of detection mean and standard deviation of multi-feature detection algorithms in different scenarios