TITLE OF THE ARTICLE



3D Shape Measurement Influencing Factors

|Stjepan Jecić |Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb, pp 102, 10002 |

| |Zagreb, Croatia |

|Nenad Drvar |Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb, pp 102, 10002 |

| |Zagreb, Croatia |

|Key words: |Measurement influencing factors, comparison, 3D scanning, structured light, laser, CMM |

|Abstract: |This paper presents a comparison of nondestructive measurement methods which are suitable for determination of |

| |the complete or partial shape and deformations of real mechanical structures. First part of this paper presents a|

| |short overview and the major differences between the existing measurement methods together with theoretical |

| |principles of sensors currently available in Croatia. The comparison was based on the analysis of the active |

| |optical projection based methods that define relative position of the measurement point by means of coherent and |

| |non-coherent structured light projection, as well as the active contact methods that define measurement point |

| |position by means of a direct physical contact with the surface of measurement object. For each of mentioned |

| |measurement methods internal and external influencing factors were defined. It is shown that the advantages and |

| |disadvantages have to be observed with respect to various aspects; like sensor types, method application, data |

| |acquisition conditions, measurement range, object reflectance, automation, accuracy, spatial resolution, method |

| |maturity, measurement planning and overall measurement costs. |

| |Second part of this paper, based upon the introduced influence factors, briefly presents the experimental |

| |verification of presented theory by comparative measurement of the actual industrial problem: saddle form surface|

| |design with burr line extraction on the raw aluminum model of a steering wheel. Measurements were conducted on |

| |the currently available digitizers: non-coherent structured light sensor "Atos", laser senzor Cyberware |

| |Mini-model Shop 3030/HIREZ/MM, and a contact probe system Ferranti Merlin 750. |

| |The achieved results justify introduced theoretical comparison based on the influence of internal and external |

| |parameters. It is shown that the wide basis of possible internal and external influencing factors require |

| |multidisciplinary approach to the comparison of each methods together with the associated measurement sensors. |

INTRODUCTION

Together with the rapid development of the electronic computers during preceding twenty years it is evidential that there has been an intensive progress in the development of contact and non-contact optical 3D shape measurement methods. During that time, not all efforts were engaged in the inventions of new measurement technologies, but were mostly dedicated to the refinement of the existing knowledge thus improving the measurement accuracy of the existing sensors. Apart from scientific work dedicated to this field, further boost to the development pace was additionally given through the availability of low cost CCD and recently CMOS sensors, cheap and eye-safe low power laser sources and various kinds of optical and mechanical components. Clearly, both hardware and software components of the measurement sensors were improved over that period of time. Since shape measurement is rather pointless without the possibility for interpretation of measurement results, there has also been a simultaneous development of numerical shape modeling methods of 3D CAD models, numerical simulation methods e.g. FEM as well as rapid prototyping methods.

Typical example of free-form surfaces that respond to high esthetical, ergonomical and technical design is the automotive industry where product design changes on a daily basis. Since manual surface modeling requires a huge effort of time and money, this motivates further development of 3D optical shape measurement techniques. Consequently, 3D measurement methods are rapidly gaining importance as industry raises its demands in high technical performance of final products, short production times, low manufacturing costs and the overall product quality. However, not only industry motivates this development, as modern medicine, heritage, architecture and other end users recognize potentials of 3D sensors. Variety of measurement problems leads to many different specialized types of 3D scanners(1) that are developed in conjunction with the actual and often very specific requirements:

❑ achievable accuracy and reliability of measurement information acquisition,

❑ requests of Zahtijevanuspatial and measurement resolution,

❑ Brzinu measurement speed and capacity,

❑ Potpunostcompleteness of shape digitalization,

❑ possibility of measurement volume modification together with the flexibility for measurement of various object shapes and sizesMogućnos,

❑ insensitivity for mechanical and optical properties of measurement object,

❑ Neosjetljivostinsensitivity for local surface shape variations,

❑ Mjernueconomical measurement and measurement charge,

❑ Kontrolureal time measurement process control,

❑ degree of automation and stiffness of the learning curve Stupanj,

❑ Mogućnostdata exchange with various CAD systems,

❑ I other specific requests.

Wide spectrum of presented requests leads to conclusion that for the optimal sensor design one has to take into account not only theoretic background with their core measurement principles but also the potentials and restrictions of the appropriate measurement sensors and its interaction with the object of measurement during the actual measurement task. Theoretical background should be sufficient for preliminary decision of which measurement method will possess characteristics sufficient for specific measurement task, but only the comprehensive knowledge of real influencing factors that characterize advantages and drawbacks of each measurement sensor over the other will lead to the selection of the supreme commercially available sensor for the specific task.

MOTIVATION

During that development period there has been an equivalent development of methods that provide accurate length measurement (distance between only two points on the object surface) and methods that provide vast number of measurement points in a relatively short period of time. Special attention was given to the development of nondestructive methods for acquisition of the complete spatial shape and deformations of measurement object whose temperature condition, geometry and position in space doesn't change during the measurement process.

Since the development of contact Coordinate Measuring Machines (CMM) that started in the beginning of 60's of last century, a substantial amount of literature has been published regarding its construction and factors influencing the reliability of measured information as well as standards that enable unique comparison and classification of the existing commercially available sensors e.g. ISO 10360-(1 to 6), ISO 14253-(1 to 4), ISO 15530-3, ANSI/ASME B89, VDI / VDE 2617-(1 to 7), VDI / VDE 2627.

Because of large amount of commercially available optical non-contact sensors whose core measurement principles are based on mutually different technologies, in a connotation of standardization, together with the unclear influence and interaction of various external and internal influencing factors – optical methods that measure the entire object's shape even after more than a quarter of a century of development still lack clear and standardized procedures for the expressing of measurement uncertainty (there are some recommendations e.g. VDI/VDE 2634 but they aren't widely recognized and accepted). The consequence of this situation is the confusing situation on the optical measurement sensor market, since the manufacturers express the expected accuracy and resolutions of their sensors based on the measurements that most of the time aren't published nor widely accepted and comparable to the others. Such situation makes it impossible to conduct periodical attesting of measurement sensors by the legal institutions, thus the users of often very expensive optical measurement equipment don't have the accurate information's regarding the current expected measurement uncertainty nor the means for suitable and reliable comparison with the rest of the commercially available sensors.

The aim of this paper is to establish the basis for a relative comparison of contact and non-contact 3D measurement methods based on the analysis of internal and external influencing paramethers that affect the accuracy and applicability of each methods and the equivalent measurement sensor.

THEORY

1 Sensor description – coded light projection sensor

The fringe projection sensor ATOS(2) produced by GOM mbH determines the position of the object point on the surface of the measurement objects via projection of visible, non-coherent structured light onto the object surface (figure 1).

[pic]

Figure 1. Schematic principle of the operation of 3D-scanner ATOS

The sensor is equipped with a light emitting projector and two cameras in convergent set-up, which provides an over-determined mathematical triangulation model. Projector purpose is just to provide the active and unique definition of object points in each camera; hence for a dual camera system projector doesn't necessarily need to be calibrated.

Fundamental problem of structured light projecting technology is in the correspondence problem, since to obtain spatial object coordinates of the spatial measurement point one needs to locate its projected image position in both cameras i.e. one has to find for each pixel in left image the corresponding pixel in the right image. Digitizing system utilized in this work overcomes the correspondence problem by employing the projection of non-coherent visible light that is structured via temporal phase shifting and Gray-code projection, as shown in figure 1.

If there is a unique relationship between the coordinates of object point image projections, for each stereo pair in both camera triangulation is conducted by bundle adjustment method.

Real objects due o their complex geometry usually cannot be digitised from a single view, therefore in order to measure the entire geometry of complex objects and thus extending the sensor measurement volume, passive reference targets are attached to the object surface or it’s rigid surrounding so their coordinates can be determined by means of photogrammetry. These passive targets define the object coordinate system in the particular referent relative sensor to object position.

2 Sensor description – laser light stripe projection sensor

A laser stripe profiler(3) consists of a laser source, camera, and linear motion platform, schematically illustrated in Figure 2. The laser source emits a plane of light which forms an illuminated stripe on the object’s surface. This is viewed by a camera displaced from the laser plane. In the indicated arrangement, the stripe runs roughly from top to bottom of the images. The horizontal displacement is related to surface shape. The data extracted from the image is in which column (x-direction) the stripe crosses each row. The position in the x-direction is estimated to sub-pixel accuracy using a sub-pixel image operator such as Gaussian interpolation. Given the camera intrinsic parameters and its relative location with respect to the laser plane, the position of the corresponding point in the laser stripe (i.e., the coordinates in the laser stripe plane) can be estimated by triangulation. Finally, these can be transformed into 3D coordinates by incorporating the known motion of the object relative to the laser plane.

[pic]

Figure 2. Schematic principle of the operation of light-stripe laser scanner

3 Sensor description – contact coordinate measuring machine

A Coordinate Measuring Machine(4) (CMM) (Figure 3) is actually a very precise Cartesian robot equipped with a tactile probe, and used as a 3-D digitizer. The probe, under computer control, touches a sequence of points in the surface of a physical object to be measured, and the CMM produces a stream of x, y, z coordinates of the contact points.

[pic]

Figure 3. Schematic principle of the operation of CMM

4 General definition of influencing factors

By the joint term external factors we'll hereafter describe all the influencing factors that affect the measurement process(5), and thus the quality of measured results, but who doesn't originate from the core mathematical principles that the observed method is built upon.

Regarding the source of the external influencing factrors, these factors can be categorized as:

❑ Measurement object's characteristics:

o Geometrical

o Material

o Optical characteristics

❑ Environmental influence:

o Temperature

o Vibrations

o Humidity

o Lightning conditions

o Dust and other sources of filth

❑ Technological influences:

o Definition of measurement task

o Measurement planning

o Equipment handling

❑ Software

❑ Human factor

❑ Additional data processing requests

The term internal factors reflect on the technological principles that measurement method utilizes for a measurement point definition. Since all the measurements have to be conducted on the real measurement sensors this definition will be extended to describe the influence of all the relevant factors that originate from the actual constructional sensor features which directly affects the quality and the distribution of digitized information.

Thus, the discussion of relevant internal influencing factors has to be conducted separately for both contact and non-contact measurement methods. To preserve comparison clarity, internal influencing factors have to be observed through common reference points:

❑ actual sensor hardware features,

❑ mathematical models of measurement point definition,

❑ measurement point distribution,

❑ measurement volume definition,

❑ influence of system calibration,

❑ influence of software.

1 Contact Coordinate Measuring Machines (CMM)

Main sources of internal influences on the accuracy have to be observed as(6):

❑ The influence of the mehanical CMM structure:

o linear movement accuracy,

o perpendicular movement error,

o angular error,

o axial positioning error.

❑ The influence of measurement head:

o measurement probe calibration,

o measurement head cinematics,

o influence of dynamic effects, measurement force and touch probe speed,

o measurement head angle alteration,

o multiple probe measurement head,

o measurement head replacement.

❑ Influence of statical i dynamic deformation of structural CCMM elements.

❑ Temperature influence.

2 Projection systems

Contact CMM in spite of the large number of hardware variants share the common structure elements that can be easily categorised as suggested in the previous chapter. Apart from that, projection sensors for the definition of measurement point utilize various technological principles (e.g. moveable glass with printed grating structures, LCD projectors, laser point or stripe projections with mechanical relative sensor-to-object movement). Thus, the definition of internal influencing factors was partially influenced by the projection systems utilized in this work.

Since internal influencing factors weren't so far in the accessible literature adequately systemized, they should be observed through the:

❑ Influence of the structure elements of the measurement device:

o image digitalisation,

o optical elements,

o light source,

o moveable elements.

❑ Measurement point and measurement volume definition,

❑ System calibration,

❑ Software.

EXPERIMENT

Due to the nature of this paper and the vast number of factors that influence the accuracy and applicability of measurement sensors, the complete experiment covering all the possible factor variants and their interactions deserves a deeper investigation as suggested and conducted in(7,8). It should be designed in such a way to easily compare influences of mentioned factors both locally and on the entire measurement volume, taking into account sensor types specific issues, method application, data acquisition conditions, measurement range, object reflectance, degree of automation, accuracy, spatial resolution, method maturity, measurement planning. Comparative measurement of the position of burr and object surface was conducted(7,8) on a real industrial problem, the aluminium steering wheel and the influence of internal and external influencing factors was critically observed.

CONCLUSION

Number of influencing factors, as discussed above, together with their origins and specific influences in various stages of measurement process suggest that one has to decide which sensor type and measurement technique will be used for a specific measurement task upon the characteristic requests for the post processing needs, taking into account the size of the measured object, required resolution, required accuracy, robustness, and acquisition time as well as the total cost of measurement. Based on our observations, such knowledge is still scarce, and research in the direction towards standardization of optical measurement methods will be continued. The results of experimental investigation in(5,6,8) justify the introduced theoretical comparison based on the influence of internal and external parameters. It was shown that due to the wide basis of possible internal and external influencing factors it is required to take multidisciplinary approach to the comparison of each method together with the associated measurement sensors.

REFERENCES

1) Raindrop Geomagic, 3D Scanner Report,

2) M. Gomerčić, Doprinos automatskoj obradi optičkog efekta u eksperimentalnoj analizi naprezanja, Phd thesis, FSB Zagreb, 1999.

3) A. M. McIvor, Calibration of a laser stripe profiler, Proceedings of Second International Conference on 3-D Digital Imaging and Modeling, 92-98, 1999.

4) S. N. Spitz, Dimensional Inspection Planning for Coordinate Measuring Machines, PhD thesis, University of Southern California, Department of Computer Science, 1999.

5) S. Jecić, N. Drvar, "The assessment of structured light and laser scanning methods in 3D shape measurements", Proceedings of the 4th International Congress of Croatian Society of Mechanics, 237-244, 2003.

6) V. Mudronja, Prilog istraživanju graničnih mogućnosti primjene trokoordinatnih mjernih uređaja s gledišta točnosti, Phd thesis, FSB Zagreb, 1989.

7) N. Drvar, Usporedba metoda za određivanje oblika i deformacija mehaničkih konstrukcija, Master thesis, FSB Zagreb, 2004.

8) D. Semenski, A. Bakić, N. Drvar, A. Marinov, "A new 3D scanning-aided procedure in cutting tool design", Proceedings of the 8th International design conference, 799-804, Dubrovnik 2004.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download