《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (5): 1583-1590.DOI: 10.11772/j.issn.1001-9081.2021030493
收稿日期:
2021-04-01
修回日期:
2021-05-18
接受日期:
2021-05-18
发布日期:
2022-06-11
出版日期:
2022-05-10
通讯作者:
李万禹
作者简介:
回立川(1980—),男,河北邢台人,副教授,博士,主要研究方向:电力系统运行监测基金资助:
Lichuan HUI, Wanyu LI(), Yilin CHEN
Received:
2021-04-01
Revised:
2021-05-18
Accepted:
2021-05-18
Online:
2022-06-11
Published:
2022-05-10
Contact:
Wanyu LI
About author:
HUI Lichuan, born in 1980,Ph. D.,associate professor. Hisresearch interests include power system operation monitoring.Supported by:
摘要:
电力巡线图像纹理复杂且具有视差变化,针对传统算法获取成对匹配点数量较少、配准精度较低,严重影响电力巡线无人机图像拼接效果等问题,提出了一种基于改进OANet的图像拼接算法。首先,借助加速“风”(AKAZE)算法对待拼接电力巡线图像进行粗匹配;其次,对OANet中Order-Aware模块添加挤压和激励网络(SENet),从而增强网络对局部和全局上下文信息的抓取能力,得到更精确的成对匹配点;然后,通过MPA算法配准待拼接图像;最后,借助内容压缩感知算法计算重叠区域的最佳缝合线以完成图像拼接。改进OANet相较原OANet的正确匹配点数量增加了10%左右,耗时平均增加了10 ms;与APAP算法、AANAP算法、MPA算法等配准拼接算法相比,所提算法的拼接质量最好,其待拼接图像的重叠区域的均方根误差为0,非重叠区域未发生畸变。实验结果表明,所提算法可快速、稳定地拼接电力巡线航拍图像。
中图分类号:
回立川, 李万禹, 陈艺琳. 基于Order-Aware网络内点筛选网络的电力巡线航拍图像拼接[J]. 计算机应用, 2022, 42(5): 1583-1590.
Lichuan HUI, Wanyu LI, Yilin CHEN. Power line inspection aerial image stitching based on Order-Aware network internal point screening network[J]. Journal of Computer Applications, 2022, 42(5): 1583-1590.
算法 | 图7(a) | 图7(b) | 图7(c) | 图7(d) |
---|---|---|---|---|
AKAZE | 1 269 | 935 | 3 607 | 9 509 |
VFC | 63 | 0 | 0 | 2 945 |
RANSAC | 9 | 69 | 30 | 1 983 |
GMS | 0 | 69 | 232 | 2 482 |
OANet | 66 | 147 | 251 | 3 078 |
本文算法 | 78 | 183 | 273 | 3 790 |
表1 不同算法的匹配点数量对比
Tab. 1 Comparison of number of matching points of different algorithms
算法 | 图7(a) | 图7(b) | 图7(c) | 图7(d) |
---|---|---|---|---|
AKAZE | 1 269 | 935 | 3 607 | 9 509 |
VFC | 63 | 0 | 0 | 2 945 |
RANSAC | 9 | 69 | 30 | 1 983 |
GMS | 0 | 69 | 232 | 2 482 |
OANet | 66 | 147 | 251 | 3 078 |
本文算法 | 78 | 183 | 273 | 3 790 |
算法 | 旋转15° | 旋转30° | ||
---|---|---|---|---|
匹配点 | 正确点 | 匹配点 | 正确点 | |
AKAZE | 1 269 | 723 | 1 269 | 672 |
VFC | 808 | 218 | 799 | 238 |
RANSAC | 4 | 4 | 9 | 9 |
GMS | 0 | 0 | 7 | 7 |
OANet | 70 | 64 | 48 | 41 |
本文算法 | 75 | 69 | 53 | 48 |
表2 不同算法的角度变化匹配数据对比
Tab. 2 Angle change matching data comparison of different algorithms
算法 | 旋转15° | 旋转30° | ||
---|---|---|---|---|
匹配点 | 正确点 | 匹配点 | 正确点 | |
AKAZE | 1 269 | 723 | 1 269 | 672 |
VFC | 808 | 218 | 799 | 238 |
RANSAC | 4 | 4 | 9 | 9 |
GMS | 0 | 0 | 7 | 7 |
OANet | 70 | 64 | 48 | 41 |
本文算法 | 75 | 69 | 53 | 48 |
算法 | 仿射变化第一组 | 仿射变化第二组 | ||
---|---|---|---|---|
匹配点 | 正确点 | 匹配点 | 正确点 | |
AKAZE | 1 269 | 836 | 1 269 | 791 |
VFC | 835 | 312 | 0 | 0 |
RANSAC | 3 | 3 | 4 | 4 |
GMS | 3 | 3 | 7 | 7 |
OANet | 71 | 65 | 76 | 70 |
本文算法 | 78 | 69 | 83 | 80 |
表3 不同算法的仿射变化匹配数据对比
Tab. 3 Affine change matching data comparison ofdifferent algorithms
算法 | 仿射变化第一组 | 仿射变化第二组 | ||
---|---|---|---|---|
匹配点 | 正确点 | 匹配点 | 正确点 | |
AKAZE | 1 269 | 836 | 1 269 | 791 |
VFC | 835 | 312 | 0 | 0 |
RANSAC | 3 | 3 | 4 | 4 |
GMS | 3 | 3 | 7 | 7 |
OANet | 71 | 65 | 76 | 70 |
本文算法 | 78 | 69 | 83 | 80 |
算法 | 图7(a) | 图7(b) | 图7(c) | 图7(d) |
---|---|---|---|---|
AKAZE | 237.81 | 339.25 | 306.00 | 369.12 |
VFC | 48.73 | 43.53 | 53.37 | 54.70 |
RANSAC | 35.98 | 34.70 | 48.49 | 49.51 |
GMS | 5.33 | 6.08 | 7.34 | 6.77 |
OANet | 2 268.49 | 2 352.39 | 2 639.34 | 2 459.10 |
本文算法 | 2 279.04 | 2 368.62 | 2 643.10 | 2 469.27 |
表4 不同算法的匹配耗时对比 ( ms)
Tab. 4 Matching time consumption comparison of different algorithms
算法 | 图7(a) | 图7(b) | 图7(c) | 图7(d) |
---|---|---|---|---|
AKAZE | 237.81 | 339.25 | 306.00 | 369.12 |
VFC | 48.73 | 43.53 | 53.37 | 54.70 |
RANSAC | 35.98 | 34.70 | 48.49 | 49.51 |
GMS | 5.33 | 6.08 | 7.34 | 6.77 |
OANet | 2 268.49 | 2 352.39 | 2 639.34 | 2 459.10 |
本文算法 | 2 279.04 | 2 368.62 | 2 643.10 | 2 469.27 |
算法 | 图7(a) | 图7(b) | 图7(c) | 图7(d) |
---|---|---|---|---|
APAP | 6.41 | 7.18 | 19.93 | 6.65 |
AANAP | 6.90 | 34.87 | 23.28 | 7.32 |
MPA | 5.45 | 8.18 | 17.85 | 6.84 |
本文算法 | 0 | 0 | 0 | 0 |
表5 不同算法的配准均方根误差对比
Tab. 5 Root mean square error comparison of registration of different algorithms
算法 | 图7(a) | 图7(b) | 图7(c) | 图7(d) |
---|---|---|---|---|
APAP | 6.41 | 7.18 | 19.93 | 6.65 |
AANAP | 6.90 | 34.87 | 23.28 | 7.32 |
MPA | 5.45 | 8.18 | 17.85 | 6.84 |
本文算法 | 0 | 0 | 0 | 0 |
算法 | 图7(a) | 图7(b) | 图7(c) | 图7(d) |
---|---|---|---|---|
APAP | 3.12 | 3.94 | 2.73 | 3.41 |
AANAP | 28.95 | 29.20 | 32.56 | 34.97 |
MPA | 6.49 | 6.87 | 5.20 | 6.59 |
本文算法 | 7.75 | 7.63 | 6.80 | 7.94 |
表6 不同算法的配准耗时对比 (s)
Tab. 6 Registration time consumption comparison of different algorithms
算法 | 图7(a) | 图7(b) | 图7(c) | 图7(d) |
---|---|---|---|---|
APAP | 3.12 | 3.94 | 2.73 | 3.41 |
AANAP | 28.95 | 29.20 | 32.56 | 34.97 |
MPA | 6.49 | 6.87 | 5.20 | 6.59 |
本文算法 | 7.75 | 7.63 | 6.80 | 7.94 |
1 | 郭一江,王卫红,郑洁,等.基于无人机巡线数据的电力走廊可视化研究[J].西南科技大学学报,2020,35(3):92-96. 10.3969/j.issn.1671-8755.2020.03.014 |
GUO Y J, WANG W H, ZHENG J, et al. Research on power corridor visualization based on UAV line inspection data [J]. Journal of Southwest University of Science and Technology, 2020, 35(3): 92-96. 10.3969/j.issn.1671-8755.2020.03.014 | |
2 | 郑贵林,张丽.自旋翼飞机电力巡线技术研究与应用[J].中国电力,2014,47(7):26-31. 10.3969/j.issn.1004-9649.2014.07.006 |
ZHENG G L, ZHANG L. Research and application of auto-gyro power line inspection technology [J]. Electric Power, 2014, 47(7): 26-31. 10.3969/j.issn.1004-9649.2014.07.006 | |
3 | 罗昊,苏盛,杨浩,等.基于FPGA的电力巡线无人机硬件加密通信方法[J].中国电力,2019,52(7):11-16. 10.11930/j.issn.1004-9649.201903080 |
LUO H, SU S, YANG H, et al. FPGA-based hardware encryption of power line patrol drones [J]. Electric Power, 2019, 52(7): 11-16. 10.11930/j.issn.1004-9649.201903080 | |
4 | 朱庆辉,尚媛园,邵珠宏,等.局部特征及视觉一致性的柱面全景拼接算法[J].中国图象图形学报,2016,21(11):1523-1529. 10.11834/jig.20161112 |
ZHU Q H, SHANG Y Y, SHAO Z H, et al. Cylindrical panorama stitching algorithm based on local features and vision consistence [J]. Journal of Image and Graphics, 2016, 21(11): 1523-1529. 10.11834/jig.20161112 | |
5 | LOWE D G. Distinctive image features from scale-invariant keypoints [J]. International Journal of Computer Vision, 2004, 60(2): 91-110. 10.1023/b:visi.0000029664.99615.94 |
6 | 韩敏,闫阔,秦国帅.基于改进KAZE的无人机航拍图像拼接算法[J].自动化学报,2019,45(2):305-314. 10.16383/j.aas.2018.c170521 |
HAN M, YAN K, QIN G S. A mosaic algorithm for UAV aerial image with improved KAZE [J]. Acta Automatica Sinica, 2019, 45(2): 305-314. 10.16383/j.aas.2018.c170521 | |
7 | ALCANTARILLA P F, BARTOLI A, DAVISON A J. KAZE features [C]// Proceedings of the 2012 European Conference Computer Vision, LNCS 7577. Berlin: Springer, 2012: 214-227. |
8 | 闫璠,张莹,高赢,等.基于AKAZE算法的图像拼接研究[J].电子测量与仪器学报,2017,31(1):36-44. 10.13382/j.jemi.2017.01.006 |
YAN F, ZHANG Y, GAO Y, et al. Research of image stitching based on AKAZE algorithm [J]. Journal of Electronic Measurement and Instrumentation, 2017, 31(1): 36-44. 10.13382/j.jemi.2017.01.006 | |
9 | ALCANTARILLA P F, NUEVO J, BARTOLI A. Fast explicit diffusion for accelerated features in nonlinear scale spaces [C]// Proceedings of the 2013 British Machine Vision Conference. Durham: BMVA Press, 2013: Article No.13. 10.5244/c.27.13 |
10 | 李振宇,田源,陈方杰,等.基于改进ORB和PROSAC的无人机航拍图像拼接算法[J].激光与光电子学进展,2019,56(23):83-91. 10.3788/lop56.231003 |
LI Z Y, TIAN Y, CHEN F J,et al. Aerial image stitching algorithm for unmanned aerial vehicles based on improved ORB and PROSAC [J]. Laser and Optoelectronics Progress, 2019, 56(23): 83-91. 10.3788/lop56.231003 | |
11 | CHUM O, MATAS J. Matching with PROSAC — progressive sample consensus [C]// Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2005: 220-226. |
12 | 牟琦,唐洋,李占利,等.基于网格运动统计算法和最佳缝合线的密集重复结构图像快速拼接方法[J].计算机应用,2020,40(1):239-244. 10.11772/j.issn.1001-9081.2019061045 |
MU Q, TANG Y, LI Z L, et al. Fast stitching method for dense repetitive structure images based on grid-based motion statistics algorithm and optimal seam [J]. Journal of Computer Applications, 2020, 40(1): 239-244. 10.11772/j.issn.1001-9081.2019061045 | |
13 | BIAN J W, LIN W Y, MATSUSHITA Y, et al. GMS: grid-based motion statistics for fast, ultra-robust feature correspondence [C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 2828-2837. 10.1109/cvpr.2017.302 |
14 | RUBLEE E, RABAUD V, KONOLIGE K, et al. ORB: an efficient alternative to SIFT or SURF [C]// Proceedings of the 2011 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2011: 2567-2571. 10.1109/iccv.2011.6126544 |
15 | 张进,赵相伟,栾吉山,等.改进FAST和对立颜色特征的向量场一致性匹配[J].测绘通报,2020(11):50-54. |
ZHANG J, ZHAO X W, LUAN J S, et al. The vector field consistent matching of FAST and opposite color features is improved [J]. Bulletin of Surveying and Mapping, 2020(11): 50-54. | |
16 | MA J Y, ZHAO J, TIAN J W, et al. Robust point matching via vector field consensus [J]. IEEE Transactions on Image Processing, 2014, 23(4): 1706-1721. 10.1109/tip.2014.2307478 |
17 | ZHANG J H, SUN D W, LUO Z X, et al. Learning two-view correspondences and geometry using Order-Aware Network [EB/OL]. [2020-02-10]. . 10.1109/iccv.2019.00594 |
18 | HU J, SHEN L, ALBANIE S, et al. Squeeze-and-excitation networks [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(8): 2011-2023. 10.1109/tpami.2019.2913372 |
19 | LIN K M, JIANG N J, LIU S C, et al. Direct photometric alignment by mesh deformation [C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 2701-2709. 10.1109/cvpr.2017.289 |
20 | 胡浩慧,倪蓉蓉,赵耀.图像内容感知缩放的检测方法研究[J].软件学报,2018,29(4):1002-1016. |
HU H H, NI R R, ZHAO Y. Research on detection method of content-aware image resizing [J]. Journal of Software, 2018, 29(4): 1002-1016. | |
21 | ZHENG Z D, WEI Y C, YANG Y. University-1652: a multi-view multi-source benchmark for drone-based geo-localization [C]// Proceedings of the 2020 28th ACM International Conference on Multimedia. New York: ACM, 2020: 1395-1403. 10.1145/3394171.3413896 |
22 | ZARAGOZA J, CHIN T J, BROWN M S, et al. As-projective-as-possible image stitching with moving DLT [C]// Proceedings of the 2013 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2013: 2339-2346. 10.1109/cvpr.2013.303 |
23 | LIN C C, PANKANTI S U, RAMAMURTHY K N, et al. Adaptive as-natural-as-possible image stitching [C]// Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2015: 1155-1163. 10.1109/cvpr.2015.7298719 |
[1] | 王怀, 王展青. 非线性约束下的准单应变换图像拼接算法[J]. 计算机应用, 2021, 41(8): 2318-2323. |
[2] | 牟琦, 唐洋, 李占利, 李洪安. 基于网格运动统计算法和最佳缝合线的密集重复结构图像快速拼接方法[J]. 计算机应用, 2020, 40(1): 239-244. |
[3] | 修春波, 马云菲, 潘肖楠. 基于距离融合的图像特征点匹配方法[J]. 计算机应用, 2019, 39(11): 3158-3162. |
[4] | 李加亮, 蒋品群. 结合变形函数和幂函数权重的图像拼接[J]. 计算机应用, 2019, 39(10): 3060-3064. |
[5] | 韩李涛, 刘海龙, 孔巧丽, 阳凡林. 基于多核计算环境的地貌晕渲并行算法[J]. 计算机应用, 2017, 37(7): 1911-1915. |
[6] | 郑继明, 苏慧嘉. 基于颜色分量间相关性的图像拼接篡改检测方法[J]. 计算机应用, 2017, 37(10): 2903-2906. |
[7] | 李鹏俊, 李建增, 宋瑶, 张岩, 杜玉龙. 基于仿真的无人机遥感图像拼接误差分析[J]. 计算机应用, 2015, 35(4): 1116-1119. |
[8] | 张旭, 高佼, 王万国, 刘俍, 张晶晶. 基于显著图的输电线路杆塔图像拼接方法[J]. 计算机应用, 2015, 35(4): 1133-1136. |
[9] | 杜丙新. 基于相对定向和小区域融合的视频图像拼接[J]. 计算机应用, 2015, 35(1): 220-223. |
[10] | 袁全桥 苏波 赵旭东 李生红. 基于高频小波子带马尔可夫特征的图像拼接检测[J]. 计算机应用, 2014, 34(5): 1477-1481. |
[11] | 王铁建 刘艳丽. 基于加速鲁棒特征的广角图像自动拼接校正算法[J]. 计算机应用, 2012, 32(09): 2576-2579. |
[12] | 吴庆双 付仲良. 一种鲁棒的序列图像自动拼接方法[J]. 计算机应用, 2012, 32(02): 499-503. |
[13] | 狄颖辰 陈云坪 陈莹莹 陈彦. 无人机图像拼接算法综述[J]. 计算机应用, 2011, 31(01): 170-174. |
[14] | 王成良 杨鹏 蔡耀泽 王振国. 孔壁全景图获取中的图像拼接技术研究[J]. 计算机应用, 2010, 30(2): 344-347. |
[15] | 李阳 林意. 改进的实时图像拼接方法[J]. 计算机应用, 2010, 30(12): 3233-3235. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||