SU Zhongyuan， KANG Jiehu， FENG Luyuan， LI Hongtong， ZHANG Zhen， SUN Zefeng， WU Bin
（State Key Laboratory of Precision Measuring Technology and Instruments， Tianjin University， Tianjin 300072， China）
Abstract： Laser stripe extraction serves to be a crucial technique in the line-structured light system, and its accuracy and speed are directly related to the measurement performance. However, the traditional Hessian matrix method may produce redundant centers and missing centers, which will limit its accuracy and robustness. Besides, the complex calculation of the method makes it difficult to be applied in real-time measurement. In order to overcome these issues and achieve real-time center extraction, an improved FPGA-friendly laser stripe center extraction method is proposed. A novel judgment function is designed to replace the maximum eigenvalue, and the numerical difference between the centers and the other domains is more salient. The center judgment criteria are modified and the non-maximum suppression is used to deal with the redundant and missing centers. Furthermore, the proposed method is implemented in FPGA to achieve real-time processing. The calculations are rationally optimized to reduce resource utilization and delay time without reducing the accuracy. The mean absolute errors are 0.003 9 pixel, 0.037 3 pixel, 0.052 0 pixel, and 0.064 6 pixel, and the root mean square deviations are 0.006 8 pixel, 0.046 9 pixel, 0.065 4 pixel, and 0.081 1 pixel, respectively, in the accuracy experiment with the noise deviations of 0, 0.01, 0.02, and 0.03. The running and delay time of the proposed method in FPGA are 14.89 ms and 216.42 μs. The experimental results verify that the proposed method is highly accurate, robust, and time-efficient.
Key words： laser stripe extraction method； Hessian matrix method； redundant and missing centers； field programmable gate array (FPGA)； line-structured light system
References
［1］LIU C, DUAN F, FU X, et al. A novel method to correct the bias of curvilinear light stripe center. Optics and Lasers in Engineering, 2022, 155: 107058.
［2］SUN Q, CHEN J, LI C. A robust method to extract a laser stripe centre based on grey level moment. Optics and Lasers in Engineering, 2015, 67: 122-127.
［3］LI Y H, ZHOU J B, HUANG F S, et al. Sub-pixel extraction of laser stripe center using an improved gray-gravity method. Sensors, 2017, 17(4): 814.
［4］TIAN Q, ZHANG X, MA Q, et al. Utilizing polygon segmentation technique to extract and optimize light stripe centerline in line-structured laser 3D scanner. Pattern Recognition, 2016, 55: 100-113.
［5］LIU X, ZHAO H, ZHAN G, et al. Rapid and automatic 3D body measurement system based on a GPU-Steger line detector. Applied Optics, 2016, 55(21): 5539-5547.
［6］DONG Z, SUN X, LIU W, et al. Measurement of free-form curved surfaces using laser triangulation. Sensors, 2018, 18(10): 3527.
［7］ZHANG L, SUN J, YIN G, et al. A cross structured light sensor and stripe segmentation method for visual tracking of a wall climbing robot. Sensors, 2015, 15(6): 13725-13751.
［8］GUO X, SHI Z, YU B, et al. 3D measurement of gears based on a line structured light sensor. Precision Engineering, 2020, 61: 160-169.
［9］QI L, ZHANG Y, ZHANG X, et al. Statistical behavior analysis and precision optimization for the laser stripe center detector based on Steger’s algorithm. Optics Express, 2013, 21(11): 13442.
［10］WANG H, WANG Y, ZHANG J, et al. Laser stripe center detection under the condition of uneven scattering metal surface for geometric measurement. IEEE Transactions on Instrumentation and Measurement, 2020, 69(5): 2182-2192.
［11］SHI X, SUN Y, LIU H, et al. Research on laser stripe characteristics and center extraction algorithm for desktop laser scanner. SN Applied Sciences, 2021, 3(3): 298.
［12］ZHANG L, ZHANG Y, CHEN B. Improving the extracting precision of stripe center for structured light measurement. Optik, 2020, 207: 163816.
［13］LIU S, ZHANG Y, ZHANG Y, et al. Research on 3D measurement model by line structure light vision. Eurasip Journal on Image and Video Processing, 2018(2018): 88.
［14］YIN X, TAO W, FENG Y, et al. Laser stripe extraction method in industrial environments utilizing self-adaptive convolution technique. Applied Optics, 2017, 56(10): 2653.
［15］HE L, WU S, WU C. Robust laser stripe extraction for three-dimensional reconstruction based on a cross-structured light sensor. Applied Optics, 2017, 56(4): 823-832.
［16］MARKOVIC＇ B R, C＇ERTIC＇ J D. Improving sub-pixel estimation of laser stripe reflection center by autoconvolution on FPGA. Journal of Circuits, Systems and Computers, 2022, 31(4): 2250063.
［17］SUN Q, LIU R, YU F. An extraction method of laser stripe centre based on Legendre moment. Optik, 2016, 127(2): 912-915.
［18］WU X, TANG N, LIU B, et al. A novel high precise laser 3D profile scanning method with flexible calibration. Optics and Lasers in Engineering, 2020, 132: 105938.
［19］PAN N, SUN J H, ZHANG X M, et al. Accurate real-Time implementation of dynamic light spot center extraction based on FPGA//2013 Third International Conference on Instrumentation, Measurement, Computer, Communication and Control, Sep. 21-23, 2013, Shenyang, China. New York: IEEE, 2013: 888-892.
［20］USAMENTIAGA R, MOLLEDA J, GARCA D F. Fast and robust laser stripe extraction for 3D reconstruction in industrial environments. Machine Vision and Applications, 2012, 23(1): 179-196.
［21］XU G, SUN L, LI X, et al. Adaptable center detection of a laser line with a normalization approach using Hessian-matrix eigenvalues. Journal of the Optical Society of Korea, 2014, 18(4): 317-329.
［22］HU Z, ZHU H, HU M, et al. Adaptive centre extraction method for structured light stripes. Ukrainian Journal of Physical Optics, 2017, 18(1): 9.
［23］XU G, ZHANG X, LI X, et al. Timed evaluation of the center extraction of a moving laser stripe on a vehicle body using the Sigmoid-Gaussian function and a tracking method. Optik, 2017, 130: 1454-1461.
［24］ZHANG Y, LIU W, WANG W, et al. Extraction method of a nonuniform auxiliary laser stripe feature for three-dimensional reconstruction of large components. Applied Optics, 2020, 59(22): 6573.
［25］COLAK S, FRESSE V, ALATA O, et al. Comparative study of laser stripe detection algorithms for embedded real-time suitability in an industrial quality control context. Journal of Physics: Conference Series, 2018, 1074: 012173.
［26］CAI H, FENG Z, HUANG Z. Centerline extraction of structured light stripe based on principal component analysis. Chinese Journal of Lasers, 2015, 42(3): 0308006.
［27］JIANG J, LIU C, LING S. An FPGA implementation for real-time edge detection. Journal of Real-Time Image Processing, 2018, 15(4): 787-797.
［28］RICHMOND D, KASTNER R, IRTURK A, et al. A FPGA design for high speed feature extraction from a compressed measurement stream//2013 23rd International Conference on Field programmable Logic and Applications, Sep. 2-4, 2013, Porto, Portugal. New York: IEEE, 2013: 13873927.
［29］LEI H, CHEN Z, YANG Z. Design and FPGA implementation of the fast algorithm for extracting laser line//2010 International Conference on Communications and Mobile Computing, Apr.12-14, 2010, Shenzhen, China. New York: IEEE, 2010: 38-42.
［30］PANG S. The research and implementation of center extraction technology of line structured light based on ZYNQ. Nanjing: Nanjing University of Posts and Telecommunications, 2022.
［31］XU G, SUN L, LI X, et al. Global calibration and equation reconstruction methods of a three dimensional curve generated from a laser plane in vision measurement. Optics Express, 2014, 22(18): 22043.
［32］STEGER C. An unbiased detector of curvilinear structures. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1998, 20(2): 113-125.
［33］ZHANG Z. A flexible new technique for camera calibration. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(11): 1330-1334.
基于FPGA的实时线结构光传感器激光条纹中心提取方法
苏中元， 康杰虎， 冯橹源， 李泓彤， 张振， 孙泽峰， 吴斌
（天津大学 精密测试技术及仪器国家重点实验室， 天津 300072）
摘要：激光条纹提取方法是线结构光传感系统中的一项重要技术， 其准确度和速度直接影响系统的测量性能。 然而， 传统Hessian矩阵法可能会产生多余中心点和缺失中心点， 限制了其准确度和鲁棒性。 此外， 传统Hessian矩阵法计算复杂， 难以应用于实时测量场景。 为此， 提出了一种适合在FPGA中实现的激光条纹中心提取改进方法。 首先， 设计了一个新的判断函数来代替最大特征值来选取显著中心点， 以增加真实中心和其他区域之间的数值差异， 并通过修改中心判断标准， 采用非最大值抑制的方法处理多余中心点和缺失中心点。 其次， 将所提出的方法在FPGA中进行实现。 在不降低精度的前提下， 对计算进行了合理优化， 减少了FPGA的资源利用率和延迟时间。 在准确度对比试验中， 在噪声水平分别为0, 0.01, 0.02和0.03时， 平均绝对误差分别为0.003 9像素, 0.037 3像素, 0.052 0像素和0.064 6像素， 均方根误差分别为0.006 8像素, 0.046 9像素, 0.065 4像素和0.081 1像素； 在FPGA中实现时， 运行时间和延迟时间分别为14.89 ms和216.42 μs。 实验结果证明， 所提出的方法准确度高、 鲁棒性好， 能够进行实时中心提取。
关键词：激光条纹中心提取方法； Hessian矩阵法； 多余中心点和缺失中心点； 现场可编程门阵列（FPGA）； 线结构光传感测量系统
引用格式：SU Zhongyuan， KANG Jiehu， FENG Luyuan， et al． A real-time laser stripe center extraction method for line-structured light system based on FPGA． Journal of Measurement Science and Instrumentation， 2023， 14（4）： 452-462. DOI： 10．3969／j．issn．1674-8042．2023．04．008
[full text view]
