基于显著矩阵与神经网络的红外与可见光图像融合 
下载： 876次
Infrared and Visible Image Fusion Based on Significant Matrix and Neural Network
兰州交通大学电子与信息工程学院, 甘肃 兰州 730070
图 & 表
图 1. NSST分解过程 [14]
Fig. 1. Decomposition process via NSST [14]
下载图片 查看原文
图 2. 算法流程图
Fig. 2. Flow chart of proposed algorithm
下载图片 查看原文
图 3. 显著性检测结果。(a)红外图像;(b) AC算法;(c) HC算法;(d)改进算法
Fig. 3. Saliency detection results. (a) Infrared image; (b) AC algorithm; (c) HC algorithm; (d) improved algorithm
下载图片 查看原文
图 4. 点火次数。(a)红外图像;(b)可见光图像;(c)差值统计
Fig. 4. Number of ignition. (a) Infrared image; (b) visible image; (c) difference statistics
下载图片 查看原文
图 5. 第一组图像。(a)红外图像;(b)可见光图像;(c) LP;(d) Contourlet;(e) NSCT;(f) Curvelet;(g) DTCWT;(h) CBF;(i) CSR;(j) JSR;(k) JSRSD;(l)所提方法
Fig. 5. First group of images. (a) Infrared image; (b) visible image; (c) LP; (d) Contourlet; (e) NSCT; (f) Curvelet; (g) DTCWT; (h) CBF; (i) CSR; (j) JSR; (k) JSRSD; (l) proposed method
下载图片 查看原文
图 6. 第二组图像。(a)红外图像;(b)可见光图像;(c) LP;(d) Contourlet;(e) NSCT;(f) Curvelet;(g) DTCWT;(h) CBF;(i) CSR;(j) JSR;(k) JSRSD;(l)所提方法
Fig. 6. Second group of images. (a) Infrared image; (b) visible image; (c) LP; (d) Contourlet; (e) NSCT; (f) Curvelet; (g) DTCWT; (h) CBF; (i) CSR; (j) JSR; (k) JSRSD; (l) proposed method
下载图片 查看原文
图 7. 第三组图像。(a)红外图像;(b)可见光图像;(c) LP;(d) Contourlet;(e) NSCT;(f) Curvelet;(g) DTCWT;(h) CBF;(i) CSR;(j) JSR;(k) JSRSD;(l)所提方法
Fig. 7. Third group of images. (a) Infrared image; (b) visible image; (c) LP; (d) Contourlet; (e) NSCT; (f) Curvelet; (g) DTCWT; (h) CBF; (i) CSR; (j) JSR; (k) JSRSD; (l) proposed method
下载图片 查看原文
图 8. 第四组图像。(a)红外图像;(b)可见光图像;(c) LP;(d) Contourlet;(e) NSCT;(f) Curvelet;(g) DTCWT;(h) CBF;(i) CSR;(j) JSR;(k) JSRSD;(l)所提方法
Fig. 8. Fourth group of images. (a) Infrared image; (b) visible Image; (c) LP; (d) Contourlet; (e)NSCT; (f) Curvelet; (g) DTCWT; (h) CBF; (i) CSR; (j) JSR; (k) JSRSD; (l) proposed method
下载图片 查看原文
表 1第一组图像的评价指标
Table1. Evaluation indicators for first group of images
| Method | IE | SF | SD | PSNR | CC |
|---|
| LP | 6.9226 | 17.6749 | 36.7475 | 38.1069 | 0.8708 | | Contourlet | 5.9524 | 13.8058 | 21.7361 | 39.4505 | 0.8733 | | NSCT | 6.8784 | 18.7416 | 37.8630 | 37.3155 | 0.9589 | | Curvelet | 6.4513 | 18.0400 | 30.1752 | 47.4130 | 0.8757 | | DTCWT | 6.9711 | 17.9428 | 37.6861 | 39.3814 | 0.9567 | | CBF | 6.7097 | 18.2130 | 36.2435 | 15.6362 | 0.7378 | | CSR | 6.9962 | 18.8367 | 39.0672 | 14.4927 | 0.7084 | | JSR | 6.7898 | 17.6050 | 37.0912 | 15.7784 | 0.7230 | | JSRSD | 6.8043 | 18.0759 | 36.4776 | 16.0170 | 0.7231 | | Proposed method | 7.0152 | 18.5225 | 39.9878 | 44.8938 | 0.9673 |
|
查看原文
表 2第二组图像的评价指标
Table2. Evaluation indicators for second group of images
| Method | IE | SF | SD | PSNR | CC |
|---|
| LP | 7.0466 | 16.8403 | 42.3188 | 28.0657 | 0.8871 | | Contourlet | 6.5824 | 13.2216 | 25.0162 | 26.6001 | 0.8964 | | NSCT | 7.0283 | 17.8264 | 41.2404 | 25.3340 | 0.9287 | | Curvelet | 7.3744 | 17.9490 | 44.5483 | 24.2827 | 0.8820 | | DTCWT | 6.7691 | 17.0669 | 29.0637 | 26.4425 | 0.8959 | | CBF | 7.1189 | 18.0580 | 37.1697 | 15.9314 | 0.8518 | | CSR | 7.0242 | 20.0650 | 43.5427 | 15.5910 | 0.8355 | | JSR | 6.2253 | 14.6749 | 37.0837 | 10.3273 | 0.9021 | | JSRSD | 6.9536 | 17.2091 | 43.6872 | 13.6388 | 0.8918 | | Proposed method | 7.1443 | 20.0895 | 44.4617 | 28.1845 | 0.9142 |
|
查看原文
表 3第三组图像的评价指标
Table3. Evaluation indicators for third group of images
| Method | IE | SF | SD | PSNR | CC |
|---|
| LP | 6.6797 | 12.5512 | 29.6264 | 36.3555 | 0.9514 | | Contourlet | 6.2594 | 8.7844 | 23.0264 | 32.8841 | 0.9628 | | NSCT | 7.0536 | 12.3911 | 35.6957 | 35.2440 | 0.9163 | | Curvelet | 7.0331 | 12.3194 | 35.7016 | 36.1933 | 0.9299 | | DTCWT | 6.9725 | 12.1452 | 35.0457 | 35.1365 | 0.9465 | | CBF | 6.4504 | 12.1774 | 26.0156 | 18.7600 | 0.9524 | | CSR | 7.0012 | 15.6992 | 33.5568 | 16.7124 | 0.9264 | | JSR | 6.9058 | 13.3103 | 35.3446 | 12.5880 | 0.8820 | | JSRSD | 6.8618 | 16.5703 | 32.7596 | 15.0245 | 0.9178 | | Proposed method | 7.1454 | 16.4353 | 42.9358 | 43.1313 | 0.9660 |
|
查看原文
表 4第四组图像的评价指标
Table4. Evaluation indicators for fourth group of images
| Method | IE | SF | SD | PSNR | CC |
|---|
| LP | 6.4739 | 15.7264 | 23.8392 | 34.0494 | 0.8321 | | Contourlet | 6.1895 | 14.3440 | 19.7070 | 40.1515 | 0.8523 | | NSCT | 7.1166 | 20.6708 | 39.0891 | 39.6542 | 0.8642 | | Curvelet | 6.7883 | 16.1749 | 34.2253 | 39.4174 | 0.8887 | | DTCWT | 6.6119 | 18.5023 | 39.7758 | 40.3779 | 0.8481 | | CBF | 6.6335 | 15.7376 | 30.0562 | 19.7635 | 0.8447 | | CSR | 7.0584 | 17.5721 | 39.6054 | 21.4555 | 0.8176 | | JSR | 6.9219 | 17.8153 | 34.9614 | 14.3536 | 0.8689 | | JSRSD | 6.8386 | 21.2222 | 32.3196 | 15.2293 | 0.8730 | | Proposed method | 7.1243 | 28.1589 | 39.9369 | 40.7522 | 0.8795 |
|
查看原文
沈瑜, 陈小朋, 苑玉彬, 王霖, 张泓国. 基于显著矩阵与神经网络的红外与可见光图像融合[J]. 激光与光电子学进展, 2020, 57(20): 201007. Yu Shen, Xiaopeng Chen, Yubin Yuan, Lin Wang, Hongguo Zhang. Infrared and Visible Image Fusion Based on Significant Matrix and Neural Network[J]. Laser & Optoelectronics Progress, 2020, 57(20): 201007.
PDF全文
