DE102011116791A1 - Method for contactless optical testing of test contours on surfaces, particularly by triangulation light section method, involves irradiating light in form of light stripe on surface in area of test contours - Google Patents

Abstract

The method involves irradiating light in the form of light stripe on the surface in area of the test contours. The light stripe extends transverse to test contours. The light reflected from the test contours is received as a line-shaped image by an optical sensor (12). The brightness values of the pixels of the optical sensor are selected repeatedly behind one another during exposure time of the surface. The selection of the brightness values of the individual pixels is checked whether the brightness lies within a predetermined range of values.

Description

I. Field of application

The invention relates to the non-contact, optical testing of test contours, in particular three-dimensional, longitudinal test contours, on surfaces.

II. Technical background

The determination of the shape of 3-dimensional contours is often required in industry to quality control defined defined contours.

The determination of the shape or other result data related to the surface shape, e.g. B. the volume, a test contour, z. As a survey, it is often carried out by means of the light section method.

In this case, a fan-shaped, so spread in only one plane, light beam is irradiated to the surface to be examined, there is a line of light and is observed at an angle, usually an acute angle to the irradiation direction, so that the course of the light line on the surface there Surveys can be seen by the observed image of the light line on an optical sensor then shows an increase, when the linear incident light beam fan runs over this survey. In general, the direction of the light line should not be parallel to the direction of the survey, but transverse to it.

Such individual images of lines of light on the object can be made many times in short time intervals while the object is moving relatively and transversely to the light line, so that the three-dimensional surface design can be determined by placing these individual height sections, ie scans, in succession. and / or related parameters such as height, volume, width, location of the surveys, etc.

Also, welds or solder joints between adjacent sheet metal parts in the automotive industry are examined in this way on their shape and thus quality, especially inadmissible depressions in the solder seam to be found before painting.

The inspection for defects is usually carried out not only under a single, but under two different observation angles.

The problem arises that - depending on the reflection behavior of the corresponding location of the surface - the image z. B. the light line on the optical sensor in certain sections over or underexposed and is not usable.

In this context, it is already known to expose the surface with different amounts of light, and to read from the individual exposure steps respectively those parts of the image of the light line on the optical sensor and store, which are correctly exposed, so are not overexposed and not underexposed.

In this way, partial images of z. B. light lines, which then back to an overall z. B. the light line must be assembled, but this is possible by means of the electronic evaluation.

The disadvantage is that for the exposure with different amounts of light not only different lighting processes, but also a respective different recording with the optical sensor must be performed, which can double the time for scanning a test contour.

III. Presentation of the invention

a) Technical task

It is therefore the object of the invention to provide a method and a device with the aid of which the optical testing can be carried out faster than with the previous approaches.

b) Solution of the task

This object is solved by the features of claims 1 and 3. Advantageous embodiments will be apparent from the dependent claims.

Starting point for the invention is the consideration that depending on the exposure - ie the product of illuminance and illumination time - the surface usually only parts of the light line are properly exposed, so neither under nor overexposed, while the other parts in each case a different exposure correct are exposed.

In order to obtain the entire light line correctly as an image, two different approaches are proposed according to the invention:
The first approach is based on the consideration that the readout of the brightness values of a single pixel or even a line from the optical sensor currently requires about 2 ms, while a typical exposure time for the surface of a object to be tested is much longer, z. B. 500 ms. Even at 50 pixels per line of light, it is thus possible to read out the values of all pixels of the light line once within 100 ms and thus perform such a read-out process within the exposure time of 500 ms even five times in succession.

Thus one can read-out at _{^{1/5, 2/5, 3/5, 4/5,}} as well as perform the full exposure of the surface of the object each time, and it is assumed that one of the exposure steps a usable image or at least partial Image of the light line results.

At each exposure stage, this usable partial image of the light line is stored in a memory.

These partial images are then automatically reassembled from memory into an overall image of the light line.

The second approach is to read the areas that are correctly exposed on the optical sensor, that is to say pixels, of the light line not at all immediately into a memory, but instead to fix it with the correct exposure value on the optical sensor (the so-called "freezing") the subsequent exposure levels these pixels no longer to expose.

For this purpose, an optical sensor is required, on the one hand-optimally for each individual pixel or at least for pixel areas-to check whether the brightness value of the pixel is within a predetermined range and, as a second step, these pixels or pixel areas can be frozen as described above , so on further exposure of the surface no more reflected light from there more.

After correctly exposing the entire optical sensor in this way, it is read out and its information content is further processed as previously known.

In both approaches, it is preferably also detected which pixel has received its brightness value that has been stored by means of which exposure level, that is to say illumination quantity of the surface.

This can be taken into account in the further processing of the brightness values of the pixels of the sensor and allows on the one hand a more accurate evaluation and on the other hand possible conclusions about the structure of the illuminated surface and the amount of light to be used in the future.

c) embodiments

Embodiments according to the invention are described in more detail below by way of example. Show it:

1 : The content of the sensor after different exposure times,

2 : the accumulated content of the optical sensor at different exposure times and

3 : the transfer from the optical sensor to the memory.

If on a surface to be checked a z. B. sublime line-shaped test contour such as a weld is checked by the light-cut triangulation method, a fan-shaped laser light beam is irradiated to the test contour by means of a laser light source so that the light line resulting on the surface is transverse to the longitudinal direction of the test contour.

The laser light reflected from the test contour becomes a linear image 4 on an optical sensor 12 , usually a CCD chip, recorded and there gives a bright line-shaped image, which in the 1 and 2 is shown inversely as a black image.

By a different viewing angle of the test contour with respect to the angle of incidence of the laser light, there is a defined geometric relationship, by means of which the shape of the image 4 the light line 3 on the optical sensor 12 can be recalculated to the cross-sectional shape of the test contour at the irradiated point and their shape can be checked.

In the present case, an upwardly curved test contour and accordingly an image representing a bulge 4 on the optical sensor 12 given.

The scanning of the test contour is carried out by a plurality of offset in the longitudinal direction of the test contour radiated light lines and their evaluation, preferably in such a small distance that a complete review of the test contour takes place. In the 1 is only one such image 4 shown.

The problem becomes clear that also in the course of the light line and thus the resulting image 4 due to inclination, surface structure, etc., the reflection ratios on the surface of the test contour are different, so that a blazed with a certain exposure line of light on the optical sensor 12 not always one consistently well recognizable image 4 results - as with the thin edge line of the image 4 implied - but actually only partial areas 4a . 4b . 4c . 4d - depending on the amount of exposure - on the optical sensor 12 correctly exposed are recognizable and thus evaluable.

A typical exposure time when irradiating a light line 3 on the surface 2 is for example 500 ms.

It may be that already after 100 ms exposure time on the optical sensor 12 the subareas 4a of the image 4 the light line are correctly exposed, as in 1a shown while the rest of the light line on the optical sensor 12 after not recognizable, since underexposed.

After an exposure time of 200 ms, for example, the partial areas 4b according to 1b on the optical sensor 12 shown correctly exposed, the remaining areas, however, not visible or not evaluable, since the sub-areas 4a overexposed and the subareas 4c and 4d lack of sufficient exposure are not yet visible.

The same applies z. B. after 300 ms exposure time according to 1c for the subareas 4c and after 400 ms according to 1d for the subarea 4d , in which then all other sections 4a - 4c overexposed appear and only partial area 4d as correctly exposed on the optical sensor 12 is available.

First of all, it must be made clear that according to the detailed representations top left in 1a such an optical sensor 12 is composed of a grid of pixels P which can be read out line by line, column by column or even individually, depending on the design of the optical sensor, and are defined in terms of their membership in a line and a row as, for example, P 2.1 = line 2, Row 1.

In order to solve this problem, two approaches are proposed according to the invention:
The first approach is in the 2 and is referred to as freezing pixels.

This requires the optical sensor 12 and / or the downstream electronic processing unit 11 be formed accordingly:
In this case, during the total exposure time, for example at fixed exposure times, in the present example, for example after 100, 200, 300 and 400 ms, it is checked in each case whether there are pixels on the optical sensor 12 are within a predetermined evaluable range of brightness, so correctly exposed and are neither overexposed nor underexposed. If such pixels exist, these pixels will be on the sensor 12 switched inactive, so their brightness level achieved in this exposure level in the optical sensor itself fixed and not further changed, despite the progressive exposure and the increasingly more on the optical sensor 12 impinging amount of light.

In the case of exposure after 100 ms, these are, for example, the pixels P and those of the subregions 4a according to 2a are affected.

At the second check time, for example after 200 ms of exposure, this check is made again, but not for those pixels that were already "frozen", ie the subregions 4a ,

At the second exposure time level after, for example, 200 ms, these could be the subregions 4b of the image 4 the light line whose pixels are now - in addition to the pixels of the subregions 4a - additionally on the optical sensor 12 Be "frozen".

After a further exposure time, then after a total of, for example, 300 ms, the same check is performed, but again without the already frozen pixels of the subregions 4a and 4b , and then the pixels in the subregions 4c detected as correctly exposed and additionally on the sensor 12 frozen, as in 2c shown.

After a further partial exposure time of another 100 ms, so z. After 400 ms, the same check is performed again, but not for the already frozen pixels of the areas 4a - 4c , and then the subarea results 4d as correctly exposed and these pixels are additionally frozen.

After the end of the total exposure time of z. B. 500 ms then becomes the complete content of the optical sensor 12 read out, then one in the direction 8th of the image 4 consistently correctly exposed image 4 should contain, which is evaluable.

It is possible that the optical sensor 12 has a design in which the verification of the brightness value of individual pixels is not possible, but the verification of defined groups of pixels. These may be two-dimensionally arranged groups, or groups in a row or column of pixels.

For example, each row can be subdivided into groups of five pixels each, which can only be checked jointly for their brightness value, then, for example, the sum of their brightness values.

If the correct total brightness of such a group of pixels is determined during one of the checks, this entire group of pixels is frozen in terms of their brightness value.

1a at the bottom left shows such a group of pixels, which do not all have the same brightness value, but this usually increases from the edge pixels of the group to the middle pixels, as shown in two exemplary curves, but is preferably of the sensor in terms of overall brightness 12 even summed up.

Because the image 4 usually a width across its course 8th of significantly more than one pixel, even such a group-wise checking and freezing is a useful solution, which, although lower in resolution, but brings a much faster review due to the lower number of brightness values to be checked as an advantage.

If the sensor 12 is not able to freeze the brightness values of individual pixels or groups of pixels, instead the multiple readout method is chosen, as in US Pat 3 shown:
In this case, the brightness values in the sensor are not checked at the defined exposure times, since the sensor does not permit this with respect to its design, but rather the sensor is read into a memory with regard to the brightness values of all its pixels 10 as in the 3a and 3b shown.

Here are of each transverse to the direction 8th the light line 4 lying row or column of pixels P, in this case the rows R1, R2, R3, etc. of the sensor 12 the brightness values H are read out and stored.

This can be done in the simplest case by the memory 10 has as many memory locations as the optical sensor 12 Pixel P has and for each pixel P of the optical sensor, both its position and its brightness value is stored, resulting in an amount of information at least equal to twice the number of pixels P.

In order to reduce the memory requirement, a variant is that, in this case row R only the position of the top by the light line 4 to store exposed pixels P as position, in this case the position of the pixel 7.1.

In the same way in the form of a rectangular grid in fact arranged in rows and rows or otherwise sorted memory 10 Then, in the same row below or behind only the brightness values H of the individual pixels P are stored, that of the image 4 are affected, beginning with the brightness value of the first pixel of that row whose position has been stored. In the row R1, these are the brightness values of pixels 7.1 to 10.1.

As can be seen, in the further course, from the fifth column through the curvature of the image 4 already touched pixels in the sixth line.

After reading the contents of the sensor 12 in the store 10 or even during readout, the brightness values are checked to see if they are within a predetermined range of brightness values. If so, they will be in memory 10 stored or remain stored there, if no, they are deleted from the memory. The correct brightness values once stored in the memory are blocked for the current exposure process and can thus no longer be superimposed during this ongoing imaging process.

At the first exposure time level at z. B. 100 ms, these become the pixels of the area 4a Its in the partial view of the sensor 12 in 3a hatched are shown.

The same readout is at the specified split times of the total exposure time, ie z. B. at 200, 300, 400 and 500 ms, also performed, which then successively the pixels of the areas 4b . 4c . 4d according to the 1 in each case additionally in the memory permanently 10 be stored at the end of the exposure thus a consistently complete record of the image 4 the light line contains as memory content, which can be further evaluated.

In the case of this method, too, instead of the brightness values H of individual correctly exposed pixels, it is possible to check only the total brightness values of defined groups of pixels or memory locations, analogously to the first method, in order, for. B. reduced the memory requirements and evaluation effort.

Preferably, especially in the second method, it is stored on which exposure stage, ie after which exposure time, the respective brightness value of the pixel or the group of pixels was stored.

This can be done in a separate field of memory, e.g. Example, per row, as well as the position of the top pixel per row is stored, or per single pixel.

LIST OF REFERENCE NUMBERS

1

2

3

surrounding area

4

distance

4a-d

subregion

5

6

detector unit

7

8th

running direction

9

 10

Storage

11

processing unit

 12

Optical sensor

13

14

light source

A

reading process

H

brightness

P

pixel

Claims (7)

Method for the contactless, optical testing of test contours ( 2 ) on surfaces, in particular by means of the triangulation-light-section method, wherein the surface in the region of the test contour ( 2 ) Light in the form of a transverse to the test contour ( 2 ) extending light line and that of the test contour ( 2 ) reflected light from at least one optical sensor ( 12 ) as a linear image ( 4 ), characterized in that a) during the exposure time of the surface, the brightness values (H) of the pixels (P) of the optical sensor ( 12 ) are read out several times in succession, b) upon each reading, the brightness values (H) of the individual pixels (P) are checked as to whether their brightness is within a predetermined value range, c) at least the pixels (P) lying within the value range, optionally with a defined environment area ( 3 ), are stored with their brightness values (H), d) during the next reading only the brightness values (H) of those pixels (P) whose brightness values (H) have not yet been stored are checked and stored at a brightness value within the predetermined value range , A method according to claim 1, characterized in that is stored during the storage, in which read-out process (A1) and / or with which exposure the storage of this brightness value (H) of the respective pixel (P) has taken place. Method for the contactless, optical testing of test contours ( 2 ) on surfaces, in particular by means of the triangulation-light-section method, wherein the surface in the region of the test contour ( 2 ) Light in the form of a transverse to the test contour ( 2 ) extending light line and that of the test contour ( 2 ) reflected light from at least one optical sensor ( 12 ) as a linear image ( 4 ), characterized in that a) during the exposure time of the surface, the brightness values (H) of the pixels (P) of the optical sensor ( 12 ) are repeatedly checked for their brightness value, b) at least the pixels (P) lying within the value range, optionally with a defined surrounding area ( 3 ) at pixels (P), in the optical sensor ( 12 ) are fixed with their previous brightness value (H) and are no longer changed in their brightness value during the exposure time (freezing), c) as the exposure proceeds, only the remaining pixels (P) of the optical sensor ( 12 d) this process is carried out several times in succession, in particular as often as the maximum exposure time of the surface allows, e) at the end of the exposure time, the brightness values (H) of the pixels ( P) of the optical sensor ( 12 ). A method according to claim 1, characterized in that in addition to reading the brightness (H) of pixels (P) on the optical sensor ( 12 ) the associated exposure of each pixel (P) is stored. Method according to one of the preceding claims, characterized in that the defined surrounding area ( 3 ) at pixels (P) only in the direction of a column or row of the optical sensor ( 12 ), in particular transversely to the direction of the image ( 4 ) of the light line, fixed and cached, in particular with frozen. Method according to one of the preceding claims, characterized in that - for the evaluation of the content of the optical sensor ( 12 ) which is taken into account with the stored exposure of each pixel (P), - test head ( 4 ) for carrying out the method according to one of the preceding claims, comprising - a light source ( 14 ), - a detector device ( 6 ) with a planar optical sensor ( 12 ) and - an electronic processing unit ( 11 ) for image data processing, Method according to one of the preceding claims, characterized in that The optical sensor ( 12 ) is capable of, for at least portions of the sensor ( 12 ), in particular each individual pixel (P), the instantaneous brightness value in the sensor ( 12 ) and only the remaining areas or pixels remain in their brightness value in the sensor ( 12 ) to change according to the changing lighting.

Images