·Table of Contents ·Methods and Instrumentation | Metrological Characteristics of 3D - Measuring Microtelevision Videoendoscopic SystemsV V Klyuev, A A Ketkovich, V N Filinov, M V FilinovMSIA "SPECTRUM" Contact |
Fig 1: Block diagram of PC-based optic television measuring defectoscopic system DX 3 DCM |
Fig 2: Optical diagram of TVH (See details in text.) |
H = Dx × m, B = D y × m | (1) |
m = Nc/n | (2) |
Fig 3: a - general principle of scale positioning; b - scale is located on pipe generatrix; c - scale is located on the circle |
H = x_{1}^{o} - x_{2}^{o} | (3) |
The given methodics is less efficient as contrasted to above-stated, however it has significant advantages in metrology.
The methodics can be distributed also on gauging of the size of located on a circle of a tube, i.e. gauging of its angular size by means of turn of a bar, on which the camera is installed, by adjustment of images of defect edges with center of a screen and reading the corresponding readouts g_{1}and g_{2} from angle gauge display. By this, Dg=g_{1}-g_{2}, and defect linear size equals to H = Dg ×CE, where CE (mm / grade) - scale interval of angle measuring scale on the bar for a particular diameter of a tube. The given methodics excludes the influence of objective distortion errors to measuring error, however it requires individual calibration of the system for any particular diameter.
For measuring of defects depth, the light section method is used in DX 3 DCM system.
In case the defect size is determined on the monitor screen, for example, height (depth) H_{scr}, this dimension is related with actual corresponding size of defect with the ratio:
H_{scr} = H × B × V | (4) |
where % is the scale of image conversion in object space, V - image scale of the television observation system which equals to:
V = V_{0} × V_{tv} | (5) |
where V_{0} is the scale of the optic system, Vtv is the scale of video channel which equals to the ratio of corresponding raster values on the display and CCD chip. Profile conversion factor equals to:
B = h/H | (6) |
i.e. to the ratio of size of the observed defect profile in the object space to its size in the normal cross-section. From the analysis it is seen that for polished surfaces B = 2 × sinb, it doesn't depend on the projection angle a, and grows with growing of the observation angle, that leads to increase of method sensitivity. However it is desirable that a » b to eliminate losses of image brightness due to incomplete use of input aperture of receiving objective a<<b. for dissipating surfaces, and only for a = b the equation is correct. Thus, at equality of projection and observation angles the profile conversion scale doesn't depend on a surface reflection factor. Also, it is essential that in case (a+b) = 90E the observation profile also lies in the plane parallel to the subject plane. Because of this it is useful to establish a = b = 45E provided there are no design limitations.
The profile conversion scale error is characterized by the value of the difference Dh = h_{t} - h_{n} where h_{t} is an actual height of the observed defect profile, h_{n} is the profile height for nominal (calculated) angles a_{n}, b_{n} and g_{H} = a_{H} + b_{H}. By targeting of TVH onto the object surface the deviations of these angles for values of Da, Db and Dg correspondingly, are possible. After the conversion we get the equation for relative error:
(7) |
The analysis displayed that by Da » Db » Dg » 3E relative error does not exceed 3%.
The given conclusion is correct, provided there is no image defocussing due to modification of the internal diameter of the object, no alteration of distance between TVH and object internal surface, etc. This causes additional error related to smoothening of the image of measuring light line, and, correspondingly, with inaccuracy of targeting the cursor on its border when measuring on halftone image.
The depth of sharpness depends on the image scale, diameter of objective inlet opening, and on an angular value of permissible diffusion spot, expressed in a radian measure, in the plane of CCD chip, which equals to e £2a/t, where a is the diagonal of CCD pixel, t is the focal distance of camera objective.
a | b | c |
Fig 4: Calibration test bodies for DX 3 DCM system: a - slot type, b - step type, c - gauging diagram. |
The measures were reseated to the basis (gauged flat massive plate made from quartz). The characteristic sizes of measures A = 10 ¸40 mms; B = 10 mm.
The reference object was disposed in front of THV so that the profile of light section is placed on center of steps, and it would be orthogonal to the plane (Fig.5.).
Fig 5: The work window of GaugePro module, with loaded image. |
where is n - number of gaugings, M - i-th value M.
The calibration was made for both channels of a system, by this the test body turned on 90° in a plane of the basis.
To research the influence of cross sectional dimensions (width) of volumetric defect to the measuring error of its depth (height) we offered the special "vertical" pattern based, as in the first case, on tip length standards. However, the standards are not placed closely to each other, but with fixed clearance which is defined by thickness Dt of a gasket between. (Fig.7).
Varying width of the gasket Dt, we estimate capability of system research metrology performances in two-parameter space (DH and Dt), i.e. functionDH =f (Dt).
The width of gaskets was changed from 0,1 up to 5 mm.
The researches was made for characteristic parameters of items defects, for diagnostic of which the system DX 3 was intended, i.e.Dt» 0,1¸
0,3 mm, DH» 0,5¸1,0 mm, by TV-camera optical objective magnification of b_{0}» 0,1^{x}(that corresponds to standard distance between objective and surface l» 100 mm and representative value of focal distance of a lens f' = 10 mm).
As established, for these conditions the defect width Dt³1 mm practically has no influence on a measuring error of its depth (height). By this, the relative error of defect profile elements gauging does not exceed . At decreasing of defect opening (up to 0,2¸0,3 mm) the error increases up to value DH»1,0 %.
The material of gaskets was taking into account maximum approximation of optical and performances of its surface (reflection factor, dispersion index, chromaticity etc.) to the applicable performances of a material of substantial object to fulfill the principle of a similarity of physical performances of object and measurement standard, adopted in a metrology. For metallic machine-building items being inspected by DX 3 DCM system, the steel polished plates with a surface roughness Ra = 0,5 microns appeared to be most suitable. For dielectric items, the gaskets made of organic glass with a different roughness of a working surface were used.
The measurement standards described above are recommended for metrology certification of systems such as DX 3 on its operation; the standards successfully have passed approbation under production conditions and are included in a package of metrology maintenance means for DX 3 DCM.
TVH dimensions (dia.Hlength), mm | 80H160 |
TV camera view field angles: | panoramic 180E, lateral 70E |
CCD chips, pixels | 2H(597H537) |
Spectral range of TV camera,Fm | b/w: 0,4)1,2; color: 0,38 ) 0,78 |
Dynamic range of TV camera, dB | >46 |
Minimal illumination intensity on the object, lux | 1 |
Object dimensions measuring error | on surface 1%, in depth 3% |
Video signal | PAL composite and Y/C, 1 Vpp |
Digital video image size, bit | 768H576H24 |
Supported video formats | TIFF (LZW, nLZW), PCX, JPEG(by compression factor 3¸ 5) |
PC Type / OS | Pentium III 500 MHz / Win98 |
Table : General technical data of basic set of DX 3 DCM |
© AIPnD , created by NDT.net | |Home| |Top| |