Traitement en cours

Veuillez attendre...

Paramétrages

Paramétrages

Aller à Demande

1. WO1997043733 - SYSTEME D'EVALUATION D'UN MODELE POLYCHROME POUR DES OBJETS ORIENTES DE MANIERE ALEATOIRE

Note: Texte fondé sur des processus automatiques de reconnaissance optique de caractères. Seule la version PDF a une valeur juridique

[ EN ]

What is claimed is:

1. A method of inspecting a multicolor pattern on an item with a processor and an imaging device, comprising:
(a) moving the item by the imaging device to generate a 5 number of image signals corresponding to the multicolor pattern;
(b) correlating each of a number of regions of the multicolor pattern to one of a number of reference color values with the processor using the image signals;
° (c) modeling the multicolor pattern by deriving a number of test value sets with the processor each
corresponding to one of the regions, each set having a first value representative of a first characteristic of a
corresponding region and a second value representative of a 5 second characteristic of the corresponding region; and
(d) comparing the sets to reference data with the processor to detect variation of the multicolor pattern.

2. The method of claim 1, wherein the first value is determined as a function of a color space first moment of 0 the corresponding region.

3. The method of claim 1, wherein the first value is determined as a function of a physical space first moment of the corresponding region.

4. The method of claim 1, wherein the first value is
5 determined as a function of a principal radius of gyration of the corresponding region.

5. The method of claim 1, further comprising repeating
steps (a) through (d) at least one time a second, each
repetition corresponding to inspection of a different item.

6. The method of claim 1, further comprising a segmented image of the multicolor pattern.

7. The method of claim 1, further comprising:
(e) establishing a first set of color values with the processor as a function of a number of color space
coordinates ;
(f) determining a second set of color values with the processor from a sample item as a function of the
coordinates ;
(g) generating the reference color values as a function of a corresponding one of a number of color pairs, each pair having a first member from the first set and a second member from the second set, the first member being closer in color space to the second member than any other color of the second set, the second member being closer in color space to the first member than any other color of the first set.

8. A system for inspecting a multicolor pattern on an item with a number of model data sets, comprising:
a conveyor configured to convey the item;
an imaging device positioned along said conveyor, said imaging device being configured to provide a number of image signals corresponding to the multicolor pattern of the item moving relative to said imaging device on said conveyor;
a processor with a memory, said processor being
responsive to said image signals to store a number of pixels in said memory corresponding to an image of the multicolor pattern, each of said pixels having a color value;
said processor being configured to correlate the color value of each of said pixels to one of a number of model color values to identify a number of pixel groups with a uniform model color value;
said processor being configured to derive a number of test value sets each corresponding to one of said groups, each set having a number of values each corresponding to a different characteristic of a corresponding group, said test value sets being sized to occupy less memory space than said 5 """pixels and being independent of orientation of the
multicolor pattern;
said processor being configured to generate an output signal corresponding to a multicolor pattern variation detected by comparing said test value sets to the model data ° sets; and
an output device responsive to said output signal to
indicate said defect.

9. The system of claim 8, wherein said output device
includes a display configured to present an image of the 5 multicolor pattern and an indicator to highlight said variation.

10. The system of claim 8, further comprising a sensor
coupled to said imaging device to provide a signal
corresponding to presence of the item in a field of view of o said imaging device.

11. The system of claim 8, wherein said processor includes a means for detecting a defectively colored portion of the
multicolor pattern, said portion having a color distinct
from any desired color of any part of the multicolor
5 pattern.

12. The system of claim 8, wherein said test value sets are sized to occupy no more than 5% of a memory space occupied by said pixels and at least one of said values is determined as a function of at least one of:
0 a color space first moment of the corresponding group;
a physical space centroid of the corresponding group;
an inter-centroid distance associated with the corresponding group; and
a principal radius of gyration of the corresponding group .

13. The system of claim 8, wherein said processor includes a means to rotationally correlate a randomly oriented image of the multicolor pattern to orientation of a model image of the multicolor pattern.

14. The system of claim 8, wherein said output device includes a sorter configured to separate the item from a number of other items on said conveyor in response to an output signal from said processor.

15. A method of inspection of an item with an imaging device and a processor, comprising:
(a) moving the item by the imaging device to provide an item image, the item image being randomly oriented with respect to orientation of a model image of the item;
(b) characterizing the item image by generating a number of characteristic values within the processor, each of the characteristic values corresponding to one of a number of rotational positions about an item image point, and each of the characteristic values being determined as a function of an optical property of a portion of the item image corresponding to each of the rotational positions;
(c) rotationally correlating the item image to the model image as a function of the characteristic values; and

(d) comparing the item image to the model image to evaluate the item after rotational correlation.

16. The method of claim 15, wherein the item image has a number of pixels, the optical property is determined as a function of a monochrome value associated with each of the pixels, and the characteristic values correspond to an angular histogram of the item image.

17. The method of claim 15, further comprising
translationally reorienting the item image to the model image as a function of the item image point.

18. The method of claim 15, wherein (d) includes displaying the item image with an indication of a defective portion.

19. The method of claim 15, wherein (d) includes:
(dl) translationally reorienting the item image to the model image as a function of the item image point; and
(d2) superimposing the item image and the model image.

20. The method of claim 15, wherein the characteristic values correspond to a first angular histogram of the item image, and (d) includes:
5 (dl) generating a second angular histogram of the model image; and
(d2) comparing the first and second angular histograms.

21. The method of claim 15, wherein the item image includes a multicolor pattern, and further comprising:
° (e) modeling the multicolor pattern by determining a number of test values corresponding to each region of a segmented image representation of the multicolor pattern generated from the item image; and
(f) comparing the test values to model data to detect a 5 defect of the multicolor pattern of the item.

22. A system for inspecting an item with a model image,
comprising :
a conveyor configured to convey the item;
an imaging device positioned along said conveyor, said imaging device being configured to provide a number of image signals corresponding to an observed image of the item, the observed image being randomly oriented with respect to orientation of the model image;
a processor, said processor being responsive to said image signals to provide a number of image values
corresponding to a number of rotational positions about an image point on said observed image, each of said image values being determined as a function of an optical property of a portion of said observed image at each of said
rotational positions;
said processor being configured to generate a display signal corresponding to an item image rotationally
correlated to the model image as a function of said image values; and
a display device responsive to said display signal to display said item image to evaluate the item.

23. The system of claim 22, wherein said item has a
multicolor pattern and processor includes a means for detecting a defectively colored portion of the multicolor pattern, said portion having a color distinct from any desired color of any part of the multicolor pattern.

24. The system of claim 22, wherein the item has a
multicolor pattern processor which includes a means for detecting a color shift defect of the multicolor pattern as a function of a number of model colors.

25. The system of claim 22, wherein said processor is configured to provide a defect signal corresponding to a defect of the item, and further comprising a sorter
responsive to said defect signal to sort the item from a number of non-defective items on said conveyor.

26. The method of claim 22, wherein said image signal includes a number of pixels, said optical property
corresponds to a monochrome value associated with each of the pixels, and said image values correspond to an angular histogram of the observed image.

27. The system of claim 22, wherein said item image is translationally reoriented as a function of said image point

28. A method of inspection of a multicolor pattern on an item with a processor, comprising:
(a) establishing a first set of color values with the processor as a function of a number of color space
coordinates ;
(b) determining a second set of color values with the processor from a first sample as a function of the
coordinates;
(c) generating a number of reference colors each being determined from a corresponding one of a number of color pairs, each pair having a first member from the first set and a second member from the second set, the first member being closer in color space to the second member than any other color of the second set, the second member being closer in color space to the first member than any other color of the first set; and
(d) comparing a number of detected color values corresponding to the multicolor pattern to the reference color values with the processor to evaluate the multicolor pattern on the item.

29. The method of claim 28, wherein (d) includes displaying the image with an indication of a detected defect.

30. The method of claim 28, wherein the reference color values are included in a segmentation look up table in the processor .

31. The method of claim 28, wherein the first set is
determined from a number of samples of the multicolor pattern by the processor.

32. The method of claim 28, wherein each of the reference colors is determined as a function of averaging the first member and the second member from the corresponding one of the number of color pairs.

33. The method of claim 28, further comprising:
(e) determining a third set of color values within the processor from a second sample as a function of the
coordinates; and
(f) pairing the reference color values to the third set 5 of color values to update the reference colors.

34. The method of claim 28, wherein the processor includes a model image of the multicolor pattern, and further
comprising :
(e) generating an observed image of the multicolor o pattern, the observed image being randomly oriented in relation to the model image;
(f) characterizing the observed image by determining a number of image values each corresponding to one of a number of rotational positions about an image point, each of the 5 image values being determined as a function of an optical property of a portion of the observed image at each of the rotational positions;
(g) rotationally correlating the item image to the model image as a function of the image values; and
0 (h) displaying the model image after rotational correlation.