DE102004041944A1 - Method for the three-dimensional measurement of objects of any kind - Google Patents

Abstract

Method for three-dimensional measuring of objects (1) of any kind, in particular for measuring castings, comprises the following method steps: DOLLAR A recording of the object (1) for imaging the two-dimensional contour (2) in different rotational positions to the camera, DOLLAR A calculation of sectional images from the contours (2) in the form of convex hulls (3), DOLLAR A construction of a three-dimensional data set from the juxtaposition of the individual sectional images (3) and DOLLAR A image of a three-dimensional image obtained from the three-dimensional data set.

Description

The The invention relates to a method for three-dimensional surveying of objects of any kind, in particular for measuring castings. It can For example, it is a crankshaft, which is for the purpose of quality control to be measured. The object should be in each of the interested Areas can be measured or imaged three-dimensionally.

Out In practice, different methods are three-dimensional Measuring objects known. As it is with the objects of interest here quite rarely dealing with rotationally symmetric parts, it is expensive from two-dimensional images three-dimensional information to win. It is basically required, unique reference points on the object to be measured to mark to make an assignment of the two-dimensional image points to be able to.

Of the present invention is based on the object, a generic Specify method, which is used for three-dimensional surveying and imaging objects of any kind, including castings, particularly suitable, without doing To make markings on the part itself.

The above object is achieved by a method according to claim 1. Thereafter, the following method steps are provided:
First of all, the object to be measured is imaged by means of a camera. Specifically, the two-dimensional contour of the object is imaged in different rotational positions relative to the camera. From the individual shots sectional images are calculated. Of very particular significance according to the invention is now that these sectional images are calculated in the form of convex hulls.

At It should be noted that it is the mathematically defined Figure of a convex hull around the polygon with the shortest Edge acts so that all the pixels of the object are at the edge or lie inside the polygon. Concave areas of an actual The sectional image is therefore "smoothed", i.e. "filled in" to the outside.

So in any case, out of contours, i. from the individual pictures of the object in different rotational positions, the sectional images in Shape of convex hulls calculated. From the relevant records are constructed three-dimensional records or calculated, namely from a juxtaposition of the individual sectional images or convex Shells. From the thus obtained three-dimensional data record can be create a three-dimensional representation of the object.

In Advantageously, the two-dimensional contour of the object shown in backlight, namely, to to create an optimal contrast. This is done at least largely opaque Object.

These it is the object to a translucent object, the Recording of the object done in the transmitted light process. This arises in contrast to the method mentioned above an inverse illustration.

In more advantageously, the sectional images, i. the convex ones shells slice by slice across the image, i. along the optical axis, preferably perpendicular to the axis of rotation, made. Corresponding become the three-dimensional data by lining up the piecemeal produced sectional images won.

It is of further advantage when taking a lens with at least largely parallel beam path, preferably a telecentric Lens, is used so that individual contour areas regardless the concrete distance to the lens are the same size.

Of Furthermore, it is advantageous that the three-dimensional Record captured image on a screen under any angular positions is representable. So it is, for example, possible to measure the To flip or flip object to him - under consideration the loss of information due to the formation of convex hulls - of all To be able to view pages and in any angular position.

It are now different ways to design the teaching of the present invention in an advantageous manner and further education. On the one hand to the claim 1 subordinate claims and on the other hand to the following explanation of an embodiment the method according to the invention referring to the drawing. In conjunction with the explanation of the preferred embodiment of inventive method With reference to the drawing are also generally preferred embodiments and further developments of the teaching explained. In the drawing shows

1 in a schematic view of an object to be measured, concretely a crankshaft, in a total of four angular positions to the lens,

2 sectional images formed from the two-dimensional contours in the form of convex hulls in a specific rotational position of the object,

3 in a schematic view, the principle of the formation of convex hulls, to understand the formation of the sectional images according to 2 .

4a - 4d the juxtaposition of individual sectional images or convex hulls to complete a three-dimensional data set, wherein the figures show the juxtaposition of one section, two sections, ten sections and fifty sections, and

5 a three-dimensional image of the object obtained from the three-dimensional data set.

The 1 and 5 serve to illustrate the method according to the invention, wherein it is the object to be measured to a crankshaft 1 is. 1 shows the silhouette 2 the crankshaft 1 in different rotational positions, namely in rotational positions of 0 °, 45 °, 90 ° and 135 ° according to the local sequence from top to bottom.

For presentation in 2 was the in 1 The angular position shown below is chosen, the contour of which 2 serves for the exemplary production of sectional images in the form of convex hulls. So these are convex hulls 3 or sectional views at four points of the contour 2 indicated by assignment arrows. These are sectional images 3 that are orthogonal to the image plane.

3 shows by way of example the formation of convex hulls 3 by applying the tangents 4 and connecting the resulting intersections with the contour. Concave areas 5 are graphically filled, so that in this respect a loss of information must be accepted.

The 4a to 4d show the construction of a 3D data set from the individual sectional images 3 , where in accordance with 4a a sectional view, according to 4b two sectional images, according to 4c ten sectional images and according to 4d fifty sectional images are joined together to form a 3D body.

5 finally shows the complete 3D data set, assembled into a three-dimensional representation, which was created according to the mathematical method of convex hull formation from two-dimensional contour images.

Otherwise be to avoid repetition on the general part of Description referenced.

Finally, be noted that the embodiment discussed above for discussion of the claimed method, but this is not on the embodiment limits.

Claims (6)

Method for three-dimensional measuring of objects ( 1 ) of any kind, in particular for the measurement of castings, comprising the following steps: 1 ) for imaging the two-dimensional contour ( 2 ) in different rotational positions to the camera, calculation of sectional images from the contours ( 2 ) in the form of convex hulls ( 3 ), Construction of a three-dimensional data set from the juxtaposition of the individual sectional images ( 3 ) and image of a three-dimensional image obtained from the three-dimensional data set. Method according to claim 1, characterized in that the article ( 1 ) is largely opaque and that the reception of the object ( 1 ) is done in backlight. Method according to claim 1, characterized in that the article ( 1 ) is largely translucent and that the reception of the object ( 1 ) occurs in transmitted light. Method according to one of claims 1 to 3, characterized in that the sectional images ( 3 ) slice by slice across the figure, ie along the optical axis, preferably perpendicular to the axis of rotation, be made. Method according to one of claims 1 to 4, characterized that for recording a lens with at least substantially parallel Beam path, preferably a telecentric lens used becomes. Method according to one of claims 1 to 5, characterized that the three-dimensional image under any angular positions is representable.