Image analysis – Applications – Surface texture or roughness measuring
Reexamination Certificate
1999-05-05
2002-08-20
Johns, Andrew W. (Department: 2721)
Image analysis
Applications
Surface texture or roughness measuring
C358S448000
Reexamination Certificate
active
06438256
ABSTRACT:
BACKGROUND OF THE INVENTION
The uniformity of surface appearance is a key attribute of many planar products, particularly coated paper products, such as paperboard. Coatings containing optical opacifiers, such as TiO
2
, provide hiding power and visual appearance uniformity in these products. A highly uniform appearance is desired in these products. The visual uniformity is related to both surface smoothness and coating thickness uniformity. Although the surface of a coating itself may be level, an objectionable mottled appearance may be caused by thickness variations of the coating which are typically caused by unevenness of the underlying surface that passed through the coating apparatus.
There is no objective definition of mottle. Mottle is usually evaluated by trained human operators who make subjective ratings of the surface appearance based on visual observations of the coated surface. Visual ratings by a number of human observers are typically employed to establish a uniformity scale which serves as an evaluation criterion. Performance of an automated imaging system, such as that of the present invention, may be evaluated against such a criterion.
For most coated surfaces, particularly coated paperboard surfaces, the reflectance variation to be quantified is quite small. The actual reflectance variation of a coated paperboard surface is typically less than the variation in apparent reflectance (shading) caused by nonuniformity of illumination of the surface and is sometimes even less than the nonuniformity of camera response across an image of the surface.
Since the uniformity of a typical reflectance reference standard is comparable to the uniformity of some of the paperboard samples to be evaluated, prior art background correction techniques used in image processing, such as that of U.S. Pat. No. 4,656,663, are usually inadequate. Overall lightness (average reflectance) differences that exist between the paperboard samples necessitates that the measurement of visual uniformity be independent of overall lightness. Because of these factors, the prior art methods do not produce accurate, reproducible results.
SUMMARY OF THE INVENTION
The present invention is an improved image analysis method to quantify visual appearance uniformity of the surface of substantially planar objects. The measurements resulting from the method of the present invention, which are substantially independent of both the image shading and the overall lightness differences among the objects, can be correlated with the human visual ratings to an R
2
correlation factor greater than 0.90. The improved image analysis method comprises: (a) utilizing an analog to digital converter whose dynamic range may be set to a first, full, range and set to a second, contrast enhanced, range; (b) establishing transformation factors based upon the lower and upper limits of the first range and the second range of the analog to digital converter; (c) creating a frame-averaged modified dark current image representing the response of the photodetector array in the absence of light; (d) setting the analog to digital converter to map the contrast enhanced camera voltage range to the full grey level output range; (e) illumination the surface to the object with the light source, the output of the light source being set to an initial output level; (f) creating a frame-averaged image of the surface of the object; (g) determining the average grey level in the image; (h) adjusting the illumination level of the object by adjusting the output of the light source and repeating steps (f) and (g) until the average light level reflected by the surface of the object causes an average grey level in the image of step (g) to be within a predetermined range of the midpoint of the enhanced contrast dynamic range of the analog to digital converter; (i) creating a frame-averaged image of the surface of the object; (j) creating a dark-current corrected image by subtracting the frame-averaged modified dark current image of step (c) from the frame-averaged image of the surface of step (i) on a pixel by pixel basis and storing the resulting image in the memory; (k) creating a window of a predetermined size for sampling the dark-current corrected image; (l) positioning the window at a random location within the dark-current corrected image and sampling the dark-current corrected image; (m) calculating a mean grey level within the window, and calculating the standard deviation of the grey levels within the window; (n) calculating a variability factor as the ratio of the standard deviation to the mean grey level, and storing the ratio in a table in the memory; (o) repeating steps (k)-(n) a predetermined number of times and calculating a mean variability factor as the average of the variability factors of step (n) and storing the mean variability factor in the memory.
The method of the present invention is believed to be advantageous over the prior art in several ways. The illumination level is set for each sample so that the image will have a predetermined average grey level value at the midpoint (127.5) of the dynamic range of the digitization. As a result, a fixed digitizer contract enhancement window may be used for all samples. Also, the output of the light source need only be stable over the period of time during which the image is being acquired, typically only a few seconds. The uniformity measurement is independent of overall lightness differences between samples. The enhanced contrast images are corrected for camera dark current. This substantially removes contributions of the camera dark current from the measured grey level variation across the image. Since the dark current image may be captured and stored as often as desired, the uniformity measurement is effectively insensitive to CCD photodetector dark current spatial distribution variations over time, which may be related to temperature changes or aging effects in the camera CCD or electronics.
REFERENCES:
patent: 3859537 (1975-01-01), Wolf
patent: 4656663 (1987-04-01), Jansson et al.
patent: 4931657 (1990-06-01), Houston et al.
patent: 5146351 (1992-09-01), Maehara
patent: 5343308 (1994-08-01), Johnston
patent: 5533139 (1996-07-01), Parker et al.
patent: 5909244 (1999-06-01), Waxman et al.
patent: 0164222 (1985-05-01), None
Canning, Jr. Robert Vincent
Lewittes Mark E.
Merrill Michael James
Rubin Barry
Simpson, III Thomas William
Azarian Seyed
E.I. du Pont de Nemours and Company
Johns Andrew W.
LandOfFree
Method of quantifying visual uniformity of planar objects does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method of quantifying visual uniformity of planar objects, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of quantifying visual uniformity of planar objects will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2916800