Modeling and verification of a five-axis laser scanning system

Today several new kinds of laser beam sensors appear with high resolution and accuracy and find their applications in reverse engineering and quality control. Due to the incapability of changing their orientations continuously in response to the surface fluctuation of a part to be measured, they are not competent for measuring parts with complicated structures. In this paper a five-axis laser scanning system integrated with a CMM, a laser beam sensor and a PH10 rotary head is proposed, which can measure complicated parts by frequently indexing the laser sensor approximately consistent with the normal direction of the surface. As the laser value is a 1-D data and the measured data must be given in 3-D expression in the world coordinate system, a system model for coordinate transformation is established. An “equivalent probe” approach is presented for system verification, and an iterative verifying process is adopted to eliminate the verification error caused by the inclination error of the laser sensor. Experiment study shows that the system can measure a part from any direction with an accuracy of 30 μm.     

Error Compensation for Three-Dimensional Line Laser Scanning Data

In this paper, the problem of compensation of 3D line laser scanning data for improved inspection accuracy is addressed. This problem is important, as a 3D line laser scanner is about one order less accurate than a touch probe. The approach taken is to compensate through error characterisation. In other words, a software compensation is performed, instead of a hardware compensation, which is probably more expensive. To do so, the errors associated with a 3D line laser scanning system are first characterised. With error characterisation, an empirical formula is obtained relating the errors to the influencing factors including the projected angle and the scan depth. This empirical formula is used to compensate for the digitised surface data obtained by the corresponding laser scanning system. The results show that the proposed approach effectively improves inspection accuracy.     

Cutting error measurement of flexspline gears of harmonic speed reducers using laser probes

This paper presents a new system of measuring cutting errors of flexspline (FS) gears of harmonic speed reducers for quality control of the hobbing machine used to cut the gear. Two laser displacement probes are used to replace the conventional measurement method that employs a combination of Van Kuven wires and a micrometer. The two laser beams are projected onto different tooth surfaces with a certain spacing. When the tooth of the gear is over-cut/under-cut, the tooth thickness decreases/increases, resulting in variations in the outputs of the laser probes. The cutting error, which is caused by the error of the cutting depth of the hobbing machine and the wear of the cutting tool, can then be obtained from a differential output of the probes. The influence of the positioning error when aligning the relative position between the gear and the probes can be reduced greatly through rotating the gear and performing the measurement over the entire teeth. Experimental results indicated that the measurement uncertainty was on the less than ±2 μm (3σ), which meets the quality control requirement. It was also confirmed that the influence of cutting oil on the gear surface can be eliminated by blowing off the oil using compressed air.     

Effect of slice thickness on the profile accuracy of model maker rapid prototyping measured by a circular triangulation laser probe

The objective of this research is to investigate the effect of slice thickness on the profile accuracy of the model maker (MM) rapid prototyping (RP) system, layer by layer, through non-contact laser probe measurement. A circular triangulation laser probe, model OTM-3A20, made by Wolf & Beck Co., was mounted on a coordinate measuring machine (CMM), as the non-contact sensor. An adjustment device for the laser probe was designed to minimise the cosine error caused by assembly inaccuracy. The alignment test of the measuring laser beam was carried out using a calibrated specimen. The systematic accuracy of the circular triangulation laser probe with respect to the surface roughness and the surface slope of the RP workpiece was investigated using a HP5529A laser interferometer system. The maximum error of 21/2D RP part profile accuracy can be improved from 220 μm to 131 μm, and the average error can be improved from 78 μm to 46 μm as the slice thickness changed from 0.127 mm (0.005 in.) to 0.0127 μm (0.0005 in). However, the machining time increases by about seven fold based on the experimental results. An overall error of 197 μm as measured by the laser probe is attainable using the finest slice thickness 0.0127 mm (0.0005 in.) for the 3D profile accuracy. To verify the accuracy of non-contact laser probe measurement, the 3D profile of the RP part was also measured by a CNC CMM, with good consistency.     

基于单目视觉的激光光束方向标定研究

在由激光位移传感器组成的测量系统中,激光光束的方向是一个关键参数.方位角和俯仰角对于一条激光光束是最为重要的两个参数.本文中提出一种基于单目视觉的激光光束方向测量方法.首先,将CCD相机放置于基础平面上方,保持相机光轴与基础平面接近于垂直状态,并利用误差为10μm的圆孔型标定板建立单目定位模型.然后将激光光束发生装置放置在基础平面上并保持位置固定,同时在基础平面上放置特制靶块,使激光光束可以投射到靶块斜面上并形成一个激光光斑.在基础平面上方放置的CCD相机可以清晰的采集到激光光斑、靶块斜面的图像,应用相关算法提取出光斑质心的二维图像坐标.沿激光光束方向以相等间距移动靶块,通过CCD相机采集每移动一次靶块在当前位置下的光斑、靶块图像.利用相关的转换公式,结合靶块本身固有参数,将光斑质心图像二维坐标转换为基础平面下的空间三维坐标.由于靶块的移动,会得到靶块不同位置下激光光斑质心的三维坐标,将这些三维坐标拟合成空间直线表征待测激光光束.拟合直线得俯仰角即为待测激光光束的俯仰角.实验中,应用高精度仪器对靶块参数进行测定,并使用高精度标定板标定相机内外参数建立相应的定位模型.测量精度主要通过单目视觉定位精度、光斑重心提取精度来保证.结果显示,待测光束的俯角最大误差达到0.02°,光束间夹角的最大误差为0.04°.     

A 6-degree-of-freedom measurement system for the accuracy of X-Y stages

A precision 6-degree-of-freedom measurement system has been developed for simultaneous on-line measurements of six motion errors of an X-Y stage. The system employs four laser Doppler scales and two quadrant photo detectors to detect the positions and the rotations of an optical reflection device mounted on the top of the X-Y stage. Compared to the HP5528A system, the linear positioning accuracy of the developed measurement system is better than ±0.1 μm to the range of 200 mm and the vertical straightness error is within ±1.5 μm for the measuring range of ±0.1 mm. The yaw and pitch errors are about ±1 arcsec, and the roll error is about ±3 arcsec within the range of ±50 arcsec.     

A portable three-dimensional stylus profile measuring instrument

This paper describes the development of an inexpensive, portable three-dimensional (3-D) stylus-based surface profiler with a scan range of 4.5 mm × 5.5 mm and a maximum vertical range of 150 μm (limited by signal conditioning electronic range) with a resolution of better than 0.9 μm and a range of 30 μm with a resolution of better than 30 nm at the highest vertical sensitivity. The instrument occupies a volume with dimensions of approximately 75 × 75 × 40 mm (L × B × H) with scan speeds up to 0.5 mm s⁻¹. Construction of instrument components from similar material results in a thermally balanced design with a reasonably low thermal coefficient of 0.27 μm K⁻¹ and a first-order time constant of approximately 40 min. The stylus probe sensor has a free resonant frequency of 49 Hz and a low damping ratio of 0.006. When in stationary contact with a steel surface, this increases to above the transducer bandwidth of 900 Hz. Calibration of the probe sensor is achieved through direct comparison against a standard stylus gauge. Lateral calibration of the specimen carriage position has been assessed by the measurement of standard gratings and laser interferometry. Planar errors caused by the motion of the carriage have been assessed by measuring an optical flat and inferring deviation from a perfect plane as an indication of the worst-case error, which in this case, was 0.2 μm (P-V). The design and construction of the internal datum and the portability of the instrument to facilitate in situ measurement of components are emphasized. Images illustrating the surface mapping capabilities of the profiler are presented.     

A low cost, high accuracy roundness measuring system

This work presents a new architecture of a roundness measuring system in which the roundness measuring accuracy is not dependent on motion accuracy of the rotary element. In this architecture, the influence of motion errors on roundness measurement is minimized by applying a new error separation technique developed by Horikawa et. al. [1 and 2], the improved Reversal Method — IRM. A prototype that uses: a rotary table supported by ball bearings, non-contact gap sensors and a computer system that collects and process sensor readings has been developed. Experimental results have shown that even using a rotary table supported by ball bearings, that has non-repetitive motion errors larger than 2μm, the final measurement repeatability is better than 0.3μm of peak-to-peak value. In order to ensure the same levels of accuracy of the proposed system, a traditional roundness measuring system design must use a more precise and therefore more expensive type of bearing with a motion error no larger than 0.1μm.     

激光跟踪仪测角误差的现场评价

激光跟踪仪是基于角度传感和测长技术相结合的球坐标测量系统,其长度测量采用激光干涉测长方法,可直接溯源至激光波长,因此,激光跟踪仪的长度测量精度远高于角度测量精度,相对而言,测角误差就成为评价跟踪仪测量精度的重要指标。为了对现场测量激光跟踪仪的测角误差进行快速有效地评价,采用跟踪仪多站位对空间中测量区域内若干个被测点进行测量,与传统基于角度交汇原理的多站位冗余测量不同,利用各站位所观测的高精度测长值建立误差方程,并通过测长方向的矢量位移对跟踪仪测长误差进行约束,获得被测点三维坐标在跟踪仪水平角和垂直角方向上的改正值,以此来评价激光跟踪仪的测角误差。通过Leica激光跟踪仪AT901-LR进行了多站位测角误差评价实验,在现场测量条件下,跟踪仪水平和垂直方向测角误差约为0.003 mm/m(1σ),符合跟踪仪的测量误差特性。     

Compensation for five DOF motion errors of hydrostatic feed table by utilizing actively controlled capillaries

Five DOF motion errors of a hydrostatic bearing table driven by a coreless type linear motor were compensated by utilizing actively controlled capillaries in the present research. The motion errors of horizontal translation and yaw were compensated simultaneously with the two actively controlled capillaries, and those of vertical translation, pitch and roll were compensated simultaneously with the three actively controlled capillaries. The actively controlled capillaries were designed and their gains were adjusted by conducting micro step response tests. The five DOF motion errors were measured ultra-precisely by combining three measuring methods as follows. The yaw and pitch errors were measured with a laser interferometer, and the roll error was measured by the reversal method. Then, the translational motion errors in the horizontal and vertical directions were measured by the sequential two-point method, where influence of the angular motion errors on the translational measurement was compensated by utilizing the measured angular motion errors. By utilizing the developed capillaries and the combined measuring methods and by applying the iterative control, the motion errors of the horizontal translation, the vertical translation, the yaw, the pitch and the roll were compensated simultaneously and reduced significantly from 0.16 μm, 0.18 μm, 1.96 arcsec, 2.26 arcsec and 0.14 arcsec to 0.02 μm, 0.03 μm, 0.03 arcsec, 0.07 arcsec and 0.02 arcsec, respectively. The remaining motion errors were less than the measuring repeatabilities which were ±0.02 μm for the translational motion errors and ±0.05 arcsec for the angular motion errors. The results show that the present method is not only practical but also effective to realize the ultra-precision feed motion whose accuracy is equal to the currently reachable measuring accuracy.     

A differential laser autocollimation probe for on-machine measurement

An optical probe based on the principle of differential laser autocollimation has been developed for the purpose of on-machine measurement of mirror shapes. The probe is so compact that it can be mounted on diamond lathes. It can be rotated by a stepping motor about an axis perpendicular to the optical axis of measurement, and has been used to measure mirror shapes on an apparatus that imitates the conditions of an on-machine measurement system. The probe can reduce remarkably the measurement errors due to vibrational and thermal noises that could not be avoided previously in on-machine measurement. Estimating from repeatability, accuracy is better than 0.1 μm in measurement of a parabolic curve whose depth is >1 mm and length is ≈ 100 mm.     

基于BP算法的激光扫描测量误差补偿

针对激光扫描仪测量精度相对比较低,不能满足一些高精度零件的测量要求,在研究激光测量原理的基础上,通过分析和试验验证,确定了倾斜角度和工件焦距设置位置这两个影响测量误差的主要因素。以这两个因素为输入样本,建立误差影响因素的BP神经网络模型,实现了测量误差的软件补偿。试验证明,此方法在很大的程度上对误差进行了补偿,提高了测量精度。     

Assessment of surface finish on bulk scattering materials: A comparison between optical laser stylus and mechanical stylus profilometers

The accuracy of autofocusing laser stylus measurement systems based on the compact disk pick-up technique has been discussed in recent years, in particular for surface roughness measurement in the submicron-micron range. In this article we present a thorough comparison between an optical laser stylus measurement system and a Talystep mechanical stylus profilometer. Sixteen surfaces having root mean square roughnesses from 0.002 μm to 3 μm for cut-off lengths <0.8 mm have been used in the study. Both bulk scattering (paper, white ceramic) and surface scattering (metal sheet) types of samples were included. Our study shows that the laser stylus exaggerates the surface profile, in particular on surfaces having laterally small but steep local slopes. For the paper surfaces the laser stylus data were on average off by a factor of two for a cut-off length of 0.68 mm.     

激光扫描追踪人体目标位姿的算法研究

针对人体目标位姿的追踪问题,对二维激光扫描仪的工作原理进行了归纳,对所测取数据的分析算法进行了研究,设计了一种基于激光图像扫描的新型算法,利用C语言与Matlab混合编程,通过对人体目标肩部截面进行椭圆拟合,实时测取了运动人体目标的距离、角度和方位信息,并通过Matlab GUI显示.二维激光扫描仪通过旋转的光学部件发射光脉冲,形成了二维的扫描面,以实现区域扫描及轮廓测量的功能.研究结果表明,利用激光扫描仪测取的人体目标的位置信息与真实的人体目标的位置信息相比基本吻合,距离范围为0m～4m,距离精度优于0.05 m,角度范围为240°,角度跟踪精度优于2°,方位范围为360°,方位精度最高1 °,该算法可用于需要对人体目标进行实时追踪的工程应用中,例如用于对人体目标实时跟随的移动平台或机器人中.     

Evaluation on 3D micro-ground profile accuracy of micro-pyramid-structured Si surface using an adaptive-orientation WLI measurement

According to the aperture of the objectives, surfaces with steep topographies greater than approximate 25° are difficult or unable to measure with white light interferometry. Hence, an adaptive-orientation measurement is proposed by adjusting the incidence angle from 51° to 21°. In this study, a micro-grinding with #3000 diamond wheel V-tip was employed to fabricate the micro-pyramid-structured Si surface with 142 μm in depth and 38 nm in surface roughness. The objective is to evaluate the micro-profile accuracy of micro-ground Si surface. First, the four micro-ground surfaces of micro-pyramid-structured surface were measured along the adaptive orientation with an incidence angle, respectively; then iterative closest point (ICP) matching was used to reconstruct the whole micro-ground surface with four adaptive-orientation measured point clouds; finally, 3D reconstruction error and characterized profile error were investigated. It is shown that the ICP matching with denoising and finishing is valid to register four adaptive-orientation measured point clouds for reconstructing an integrated micro-ground surface. Moreover, a decrease in incidence angle to measured surfaces leads to a decrease in 3D reconstruction error, an increase in valid top-topographic point number and a decrease in characterized profile error. It is confirmed that the adaptive-orientation measurement with 21° incidence angle may enhance 3D reconstruction accuracy by about 35%, valid top-topographic point number by about 3 times and characterized profile accuracy by about 38% against the traditional measurement, respectively. The micro-ground form error of 5.5 μm and the characterized profile error of 6.0 μm may be achieved, respectively, thus the micro-grinding is valid for the precision micro-fabrication of micro-structured surface.     

High-accuracy error-reduction method for 3D complex shape measurement with local-maximum-power-based technique of FMCW

To expand the measurement target from that containing simple cylindrical-shaped components to complex-shaped components that have several directional tilted surfaces, we determined an issue in measuring the rough surfaces with laser distance measurement. The issue is that laser speckle causes measurement error, which amount, we found, can be determined by the diameter and irradiation angle of the measurement laser beam. We then developed two error-reduction techniques that are based on the optical properties of a detected waveform and power. We conducted a basic experiment and a feasibility study with actual samples, to evaluate the distance measurement performance of targets with various surface roughness and irradiation angles. The roughness and angle range of the evaluation was determined considering actual measurement needs in parts manufacturing. We confirmed that the measurement error, whose maximum value was nearly 80 μm, decreased to less than 20 μm using our proposed local-maximum-power-based technique.     

A laser sensor with multiple detectors for freeform surface digitization

An integrated laser scanning sensor for freeform surface digitization is presented. The sensor consists of a diode laser light source and four position-sensitive device (PSD) detectors. The stand-off distance of the sensor is 180 mm and the measurable range is 90 mm. The Lambert model is applied to calculate the displacement between the sensor and the measured point on the object surface along the optical axis, under the assumption of a diffusive surface. The inclination angle of the measured point from the vertical plane of the laser beam is calculated by mathematical inference. Those data are used in error compensation to improve system precision. The new design of multiple detectors could increase the measurable angle and could solve the dead space problem in single point laser of triangulation measurement. The computer simulation and actual measurements show that the displacement resolution reaches 50 μm, and the system performs well in regards to stability and repeatability. The sensor system could be mounted on the NC machine orX-Y platform for freeform surface digitization.     

LD注入调频扫描系统研究

为了减小扫描系统测量误差,可以利用半导体激光器注入式调频装置和光栅来建立扫描系统。通过改变注入电流来改变半导体激光器的输出光频,使输出光入射在光栅上,随着频率的改变,衍射光的衍射角会发生改变,进而衍射光点的位置发生改变,从而实现无机械光点扫描,能够消除扫描测量过程中机械抖动和振动带来的误差。     

Defect characterization of roller bearing surfaces with laser Doppler vibrometry

Laser Doppler vibrometry is presented as an effective method of surface profile measurement and surface crack detection in rolling bearing balls. For both inspections, a surface velocity profile is used in addition to surface displacement profile. An experimental setup and a signal-processing algorithm are used to obtain digitized surface profiles and power spectral density measurements of ball surfaces. Profile measurements are presented that match lobed ball frequency results with a resolution of 0.025 μm. Velocity measurement provides a much higher bandwidth for defect detection. Cracks down to 22 μm in width can be detected with a 150 μm by 500 μm wide laser spot size. Cracks smaller than 22 μm can be detected by decreasing the laser spot size. The method is well suited for a manufacturing environment because of its durability, simplicity, and high speed.