-
摘要:为了识别空间目标的椭圆部件，提出了一种基于自适应光学图像的椭圆检测方法。首先，利用RL(Richardson-Lucy)方法对自适应光学图像进行复原，在此基础上，采用弧支撑线段(Arc-Support Line Segments, ASLS)方法对复原图像进行椭圆检测。针对ASLS算法使用的Canny边缘提取算法带来的“弧段过分割”和“语义信息差”等问题，提出了基于多尺度组合分组(Multiscale Combinatorial Grouping, MCG)边缘提取的解决方法。最后，针对ASLS算法使用优度指标等验证方法存在部分虚假椭圆的情况，综合利用多种几何指标进行约束，有效地消除了虚假椭圆。实验结果表明：椭圆中心点检测误差优于3 pixels，半长轴误差优于4 pixels，方向角误差优于3°。在重叠面积门限为0.65时，本文算法的准确率为85.7%、召回率为93.3%，F值指标为0.893，优于传统椭圆检测算法。Abstract:In order to identify the elliptical components of space target, an ellipse detection method based on adaptive optical image is proposed. Firstly, the RL(Richardson-Lucy) method is used to restore the adaptive optics image. Next, the Arc-Support Line Segments (ASLS) method is used to detect the ellipse of the restored image. To tackle the problems of “arc segment over segmentation” and “semantic information difference” caused by Canny edge extraction, an improved edge extraction algorithm based on Multiscale Combinatorial Grouping (MCG) is proposed. Finally, for some false ellipses produced by using verification methods such as goodness measurement, a variety of geometric constraint measurement are comprehensively used to effectively eliminate the false ellipse. The experimental results show that the detection error of ellipse center point, the semi-major axis error and the direction angle error are less than 3 pixels, 4 pixels and 3 degrees, respectively. When the overlap area threshold is 0.65, the accuracy rate of this algorithm is 85.7%, the recall rate is 93.3% and the F value is 0.893. Our method is better than the traditional ellipse detection algorithms.
-
图 6空间目标自适应光学复原图像边缘提取结果
Figure 6.Edge extraction results of adaptive optics restoration image of space targets
图 7空间目标自适应光学复原图像椭圆检测结果
Figure 7.Ellipse detection results of adaptive optics restored images of partial space targets
图 8重叠面积门限与检测指标关系
Figure 8.Relationship between overlapping area threshold and detection index
表 1仿真图像椭圆参数平均误差
Table 1.Average error of linear structure components for test
平均误差(像素) 中心
$ x $中心
$ y $方向角
$ \varphi $半长轴
$ a $半短轴
$ b $ELSDc 40.10 37.81 47.26° 44.69 51.35 AAMD 2.84 10.13 10.33° 13.43 17.28 ASLS 2.57 5.08 2.11° 6.68 4.32 本文算法 1.73 2.14 2.27° 3.82 2.17 
下载:
导出CSV
表 2算法检测指标及平均耗时
Table 2.Average consumed times of those algorithms and the error detection rates
ELSDc AAMD ASLS 本文算法 准确率(%) 28.6 51.7 69.1 85.7 召回率(%) 43.7 66.7 72.3 93.3 F值 0.466 0.641 0.707 0.893 平均耗时(s) 10.058 0.525 0.659 12.874 下载:
导出CSV
-
[1] 孙志伟, 刘伟奇, 吕博, 等. 大景深空间目标成像光学系统设计[J]. 液晶与显示,2021,36(11):1597-1604.doi:10.37188/CJLCD.2021-0183SUN ZH W, LIU W Q, LYU B,et al. Design of imaging optical system for space target with large depth of field[J].Chinese Journal of Liquid Crystals and Displays, 2021, 36(11): 1597-1604. (in Chinese)doi:10.37188/CJLCD.2021-0183 [2] 李正炜, 王建立, 吴元昊, 等. 基于单站地基望远镜的空间目标姿态估计方法[J]. 中国光学,2016,9(3):371-378.doi:10.3788/CO.20160903.0371LI ZH W, WANG J L, WU Y H,et al. Method of attitude estimation for space object based on single ground-based telescope[J].Chinese Optics, 2016, 9(3): 371-378. (in Chinese)doi:10.3788/CO.20160903.0371 [3] 张磊, 吴金灵, 刘仁虎, 等. 光学自由曲面自适应干涉检测研究新进展[J]. 中国光学,2021,14(2):227-244.doi:10.37188/CO.2020-0126ZHANG L, WU J L, LIU R H,et al. Research advances in adaptive interferometry for optical freeform surfaces[J].Chinese Optics, 2021, 14(2): 227-244. (in Chinese)doi:10.37188/CO.2020-0126 [4] LU W, TAN J L. Detection of incomplete ellipse in images with strong noise by iterative randomized Hough transform (IRHT)[J].Pattern Recognition, 2008, 41(4): 1268-1279.doi:10.1016/j.patcog.2007.09.006 [5] 李娜, 王军, 董兴法, 等. 基于改进Hough变换的指针式仪表识别方法[J]. 液晶与显示,2021,36(8):1196-1203.doi:10.37188/CJLCD.2020-0179LI N, WANG J, DONG X F,et al. Pointer meter recognition method based on improved Hough transform[J].Chinese Journal of Liquid Crystals and Displays, 2021, 36(8): 1196-1203. (in Chinese)doi:10.37188/CJLCD.2020-0179 [6] ARELLANO C, DAHYOT R. Robust ellipse detection with Gaussian mixture models[J].Pattern Recognition, 2016, 58: 12-26.doi:10.1016/j.patcog.2016.01.017 [7] 吴海滨, 魏喜盈, 刘美红, 等. 结合空洞卷积和迁移学习改进YOLOv4的X光安检危险品检测[J]. 中国光学,2021,14(6):1417-1425.doi:10.37188/CO.2021-0078WU H B, WEI X Y, LIU M H,et al. Improved YOLOv4 for dangerous goods detection in X-ray inspection combined with atrous convolution and transfer learning[J].Chinese Optics, 2021, 14(6): 1417-1425. (in Chinese)doi:10.37188/CO.2021-0078 [8] YANG T, SRIHARI S N. Ellipse detection using sampling constraints[C].Proceedings of the 2011 18th IEEE International Conference on Image Processing, IEEE, 2011: 1045-1048. [9] PĂTRĂUCEAN V, GURDJOS P, VON GIOI R G. Joint aContrarioellipse and line detection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(4): 788-802.doi:10.1109/TPAMI.2016.2558150 [10] CHEN S L, XIA R B, ZHAO J B,et al. A hybrid method for ellipse detection in industrial images[J].Pattern Recognition, 2017, 68: 82-98.doi:10.1016/j.patcog.2017.03.007 [11] MENG C, LI ZH X, BAI X ZH,et al. Arc adjacency matrix-based fast ellipse detection[J].IEEE Transactions on Image Processing, 2020, 29: 4406-4420.doi:10.1109/TIP.2020.2967601 [12] LU CH SH, XIA S Y, SHAO M,et al. Arc-support line segments revisited: an efficient high-quality ellipse detection[J].IEEE Transactions on Image Processing, 2020, 29: 768-781.doi:10.1109/TIP.2019.2934352 [13] 张帆, 韩树奎, 张立国, 等. Canny算法的GPU并行加速[J]. 中国光学,2017,10(6):737-743.doi:10.3788/co.20171006.0737ZHANG F, HAN SH K, ZHANG L G,et al. Parallel acceleration of Canny algorithm based on GPU[J].Chinese Optics, 2017, 10(6): 737-743. (in Chinese)doi:10.3788/co.20171006.0737 [14] PONT-TUSET J, ARBELÁEZ P, BARRON J T,et al. Multiscale combinatorial grouping for image segmentation and object proposal generation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(1): 128-140.doi:10.1109/TPAMI.2016.2537320 [15] 刘聪, 董文飞, 蒋克明, 等. 基于改进分水岭分割算法的致密荧光微滴识别[J]. 中国光学,2019,12(4):783-790.doi:10.3788/co.20191204.0783LIU C, DONG W F, JIANG K M,et al. Recognition of dense fluorescent droplets using an improved watershed segmentation algorithm[J].Chinese Optics, 2019, 12(4): 783-790. (in Chinese)doi:10.3788/co.20191204.0783 [16] EVERINGHAM M, MULLER H, THOMAS B. Evaluating image segmentation algorithms using the Pareto front[C].Proceedings of the 7th European Conference on Computer Vision, Springer, 2002: 34-48. [17] https://img2.baidu.com/it/u=1967449378,3214789687&fm=26&fmt=auto [OL]. [18] 赵云峰. 结合自适应核函数的Mean-shift改进算法[J]. 液晶与显示,2016,31(12):1143-1148.doi:10.3788/YJYXS20163112.1143ZHAO Y F. Improved mean-shift algorithm combined with adaptive kernel function[J].Chinese Journal of Liquid Crystals and Displays, 2016, 31(12): 1143-1148. (in Chinese)doi:10.3788/YJYXS20163112.1143 [19] PRASAD D K, LEUNG M K H, QUEK C,et al. DEB: definite error bounded tangent estimator for digital curves[J].IEEE Transactions on Image Processing, 2014, 23(10): 4297-4310.doi:10.1109/TIP.2014.2346018 [20] FISH D A, BRINICOMBE A M, PIKE E R,et al. Blind deconvolution by means of the Richardson-Lucy algorithm[J].Journal of the Optical Society of America A, 1995, 12(1): 58-65.doi:10.3969/j.issn.2095-1531.2011.05.017 [21] http://www.astrospider.com/images/Lacrosse/051113_19stack.jpg [OL]. [22] MATSON C L, BORELLI K, JEFFERIES S,et al. Fast and optimal multiframe blind deconvolution algorithm for high-resolution ground-based imaging of space objects[J].Applied Optics, 2009, 48(1): A75-A92.doi:10.1364/AO.48.000A75 [23] https://gimg2.baidu.com/image_search/src=http%3A%2F%2Fwww.lunwenstudy.com%2Fuploads%2Fallimg%2F161014%2F14-161014104I5E7.png&refer=http%3A%2F%2Fwww.lunwenstudy.com&app=2002&size=f9999,10000&q=a80&n=0&g=0n&fmt=auto?sec=1652597051&t=c00e3ce939c8727f9f68d32e171e6747 [OL]. [24] PĂTRĂUCEAN V, GURDJOS P, VON GIOI P G. A parameterless line segment and elliptical arc detector with enhanced ellipse fitting[C].Proceedings of the 12th European Conference on Computer Vision, Springer, 2012: 572-585. -
下载:
点击查看大图
图(8)/
表(2)
计量
- 文章访问数:1128
- HTML全文浏览量:453
- PDF下载量:172
- 被引次数:0
