Method and device for recognising and analysing surface defects in three-dimensional objects having a reflective surface, in particular motor vehicle bodies
A method and device recognize and analyze surface defects in three-dimensional objects having a reflective surface, in particular motor vehicle bodies. In which method the surface defects are identified by the evaluation of an image, recorded by a camera in the form of a raster image of pixels, of an illumination pattern projected by a first illumination device onto a part of the reflective surface using a two-dimensional raster coordinate system. The surface defects are identified exclusively using two-dimensional image information with the aid of image processing algorithms without the need for “environmental parameters”, and complex geometric calculations can be omitted. The solution is fast and robust and can be carried out using differently configured first illumination devices, which makes it suitable for mobile applications, for example as a hand-held module. It is also made possible for the method to be optimized by a “deep learning” strategy.
1. A method for recognizing and analyzing surface defects in three-dimensional objects having a reflective surface, which comprises the steps of:
identifying the surface defects on a basis of an evaluation of a single at least one image recorded by at least one camera in a form of a grid graphic made up of pixels of an illumination pattern projected by at least one first illumination unit on at least a part of the reflective surface on a basis of a two-dimensional grid coordinate system;
identifying a region of a recorded surface of the object, on which the illumination pattern is projected, within the at least one image recorded by the at least one camera, and associating the two-dimensional grid coordinate system having x and y coordinates for each pixel with an identified region;
within the identified region, assigning intensity values to at least a part of the pixels starting from a zero point of the two-dimensional grid coordinate system;
identifying, by means of edge detection, starting from previously defined starting values, at least sections of the illumination pattern projected on the reflective surface in the at least one image; and
identifying as associated the pixels having the intensity values that are equal or similar up to a predetermined deviation value.
2. The method according to claim 1 , which further comprises forming the illumination pattern as a stripe pattern.
3. The method according to claim 1 , wherein the at least one image in the form of the grid graphic made up of pixels of at least the part of the reflected surface of an object is recorded by means of the at least one camera:
wherein after a manual or automatic configuration of a focus of the at least one camera or multiple cameras, an optimum exposure time is determined and configured on a basis of a color of the reflective surface of the object; and/or
wherein the at least one camera or the multiple cameras are calibrated with a aid of a calibration means.
4. The method according to claim 1 , wherein on a basis of a group of pixels identified as associated, a mathematical relationship is produced for a respective group of the pixels identified as associated.
5. The method according to claim 4 , wherein for each pixel of the group of pixels identified as associated, a first derivative of the mathematical relationship is determined.
6. The method according to claim 5 , wherein at least sections of the illumination pattern projected on the reflective surface in the at least one image or images are classified on a basis of the first derivative of the mathematical relationship.
7. The method according to claim 1 , wherein the surface defects are identified within the identified region on a basis of characteristic changes and/or a disappearance of a first derivative of a found mathematical relationship in classified sections.
8. The method according to claim 1 , wherein a size of an identified individual one of the surface defects is ascertained by determining a center point of a respective surface defect and at least one extremum spaced apart from the center point.
9. The method according to claim 1 , which further comprises disposing at least one marker on the reflective surface of the object and a position of the at least one marker is determined within a recorded image or images.
10. The method according to claim 1 , wherein:
the at least one first illumination unit and/or the at least one camera is moved in relation to a three-dimensional object; or
the three-dimensional object is moved in relation to the at least one first illumination unit and/or the at least one camera.
11. The method according to claim 1 , wherein:
identified surface defects are classified by a comparison to at least one model function, and/or a pattern of already classified surface defects, characteristic changes, and/or a disappearance of a first derivative of an already found mathematical relationship, and a measurement and/or evaluation data of the surface defects classified in such a way are stored in a database; and/or
the identified surface defects are classified by a comparison to database entries and measurement and/or evaluation data of classified surface defects are stored in the database.
12. The method according to claim 1 , wherein the three-dimensional objects are motor vehicle bodies.
13. The method according to claim 2 , wherein the stripe pattern has a sinusoidal intensity curve.
14. The method according to claim 3 , wherein the optimum exposure time is determined and configured by means of a control and data analysis unit, of the at least one camera or the multiple cameras on the basis of the color of the reflective surface of the object being a color of a motor vehicle body.
15. The method according to claim 1 , wherein:
the intensity values are light-dark values;
by means of the edge detection, starting from the previously defined starting values, the at least sections of the illumination pattern projected on the reflective surface, namely stripes of a stripe pattern, are identified in the at least one image or the images; and
the pixels having the light-dark values which are equal or similar up to the predetermined deviation value are identified as associated.
16. A device for recognizing and analyzing surface defects in three-dimensional objects having a reflective surface, the device comprising:
a first illumination unit for projecting an illumination pattern on the reflective surface of an object;
a camera being activatable for recording images in a form of a grid graphic made up of pixels; and
a control and data analysis unit configured to identify the surface defects on the reflective surface of the object on a basis of at least one image recorded by means of said camera in the form of the grid graphics made up of the pixels on a basis of a two-dimensional grid coordinate system;
said control and data analysis unit additionally configured to:
identify a region of a recorded surface of the object, on which the illumination pattern is projected, within said at least one image recorded by the at least one camera, and associate said two-dimensional grid coordinate system having x and y coordinates for each pixel with an identified region;
within said identified region, assign intensity values to at least a part of the pixels starting from a zero point of said two-dimensional grid coordinate system;
identify, by means of edge detection, starting from previously defined starting values, at least sections of the illumination pattern projected on the reflective surface in said at least one image; and
identify as associated the pixels having the intensity values that are equal or similar up to a predetermined deviation value.
17. The device according to claim 16 , wherein:
said first illumination unit is configured as an illumination arch movable on at least two rollers, said first illumination unit having at least one mirror;
said at least one mirror is configured to project by reflection an illumination pattern of said first illumination unit on the reflective surface of the object; and/or
said first illumination unit contains at least one second illumination unit, and said at least one second illumination unit is configured to project an illumination pattern on the reflective surface of the object.