计算机科学 ›› 2023, Vol. 50 ›› Issue (11A): 221100170-5.doi: 10.11896/jsjkx.221100170
陈海燕1, 朱军林1, 王平2
CHEN Haiyan1, ZHU Junlin1, WANG Ping2
摘要: 利用圆形靶标进行相机标定时,靶标成像效果会随着不同的相机拍摄位姿呈现为椭圆,因此利用常规圆心定位方法得到的图像圆心坐标并非真实圆心在图像中的成像位置,直接利用该圆心图像坐标进行相机标定的标定精度不高。针对此问题,提出了一种先对圆形靶标图像圆心定位误差进行预测建模,然后进行误差补偿来提高圆心定位精度的方法。首先,建立圆形靶标成像图的仿真图像集;其次,对图像预处理并利用椭圆拟合法定位图像中的圆心坐标;再次,构建并训练GA-BP神经网络,建立圆心定位误差与相机镜头位姿之间的关系模型;最后,通过误差补偿策略对定位的圆心坐标进行误差补偿。实验结果表明,所构建的GA-BP神经网络模型对圆心定位的横、纵坐标的误差预测精度明显优于BP或者E-R模型,其MAPE,RMSE,R2分别为5.51%,0.004 8,0.999 6和6.14%,0.096 4,0.999 8。误差补偿后的圆心定位精度更高,验证了采用误差预测建模和误差补偿的方法提高圆心定位精度的可行性,为高精度相机标定任务提供了方法支撑。
中图分类号:
[1]ZHOU Q,YANG Y M,WANG Z Z.Structured light measure-ment technology based on binocular stereo vision[J].Computer Engineering,2018,44(7):244-249,258. [2]WANG G H,QIAN K M.A review of calibration methods forlinear array cameras[J].Acta Optica Sinica,2020,40(1):181-193. [3]XU Y Z,ZHAO R J,MA Y B,et al.A linear calibration method for fish eye lens[J].Computer Engineering,2019,45(12):196-200. [4]CHEN T F,MA Z,WU X.Optical plane calibration in linestructured light sensor based on active vision[J].Optics and Precision Engineering,2012,20(2):256-263. [5]PENG Y L,LIU S G,PEI Z,et al.Linear Self-calibration me-thod based on Non-rigid body trajectory basis[J].Acta Electronica Sinica,2017,45(1):135-139. [6]ZHANG Z Y.A flexible new technique for camera calibration[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2000,22(11):1330-1334. [7]PENG Y,GUO J B,YU C Q,et al.Calibration method for high precision camera based on plane transformation[J].Journal of Beijing University of Aeronautics and Astronsutics,2022,48(7):1297-1303. [8]WU H B,XU R T,WANG A L,et al.Overview of 3D reconstruction technology of human lumen based on computer vision[J].Computer Engineering,2021,47(10):1-15. [9]XING D K,DA F P,ZHANG H.Research and application ofprecision positioning method for circular target[J].Chinese Journal of Scientific Instrument,2009,30(12):2953-2958. [10]YANG S R,LIU M,YIN S B,et al.An improved method for location of concentric circles in vision measurement[J].Measurement,2017,100:243-251. [11]ZHOU F Q,WANG Y X,PENG B,et al.A novel way of under-standing for calibrating stereo vision sensor constructed by a single camera and mirrors[J].Measurement,2013,46(3):1147-1160. [12]WANG S K,ZHAO J Z,JIANG M,et al.Zhang’s camera calibration method based on circular array calibration plate[J].Transactions of Beijing Institute of Technology,2019,39(8):859-863. [13]MA C,YANG J,MEI X S,et al.Thermal error modeling ofspindle based on genetic algorithm and BP network[J].Compu-ter Integrated Manufacturing Systems,2015,21(10):2627-2636. [14]SADEGHI B.A BP-neural network predictor model for plasticinjection molding process[J].Journal of Materials Processing Technology,2000,103(3):411-416. [15]HUANG X W,CHEN S F,ZHOU T Y,et al.A survey of genetic algorithms for flexible Job-shop scheduling[J].Computer Integrated Manufacturing Systems,2022,28(2):536-551. [16]FANG C H,ZHANG F,DANG W J,et al.Recognition method of main insulation surface defects of 10 kV cable based on Canny algorithm[J].Optoelectron · Laser,2021,32(11):1188-1195. [17]FITZGIBBON A,PILU M,FISHER R B.Direct least square fitting of ellipses[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1999,21(5):476-480. [18]JIANG J W,KANG J H,WU B.Research on high-precision positioning compensation algorithm of laser spot center[J].Advances in Laser and Optoelectronics,2021,58(14):362-367. |
|