Television – Camera – system and detail – With single image scanning device supplying plural color...
Reexamination Certificate
1997-09-12
2002-05-14
Garber, Wendy R. (Department: 2612)
Television
Camera, system and detail
With single image scanning device supplying plural color...
C348S272000, C382S167000
Reexamination Certificate
active
06388706
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to an electronic camera and an image processing method, more specifically, in a single plate electronic camera employing a photoelectric conversion element and a color filter, the electronic camera wherein a color signal value of which color is different from a filter color of each pixel is obtained by interpolation, and a method of an image compression and decompression process employed in said electronic camera and the like.
Conventionally, the single board sensor electronic camera has been known wherein a color filter is put on each of pixels of a photoelectric conversion element such as a CCD (Charge Coupled Device), and color image signals are obtained.
In the above-mentioned single board sensor electronic camera, as shown, for example, in
FIG. 2
, a mosaic color filter is employed wherein in order to obtain a brightness signal (Green signal) for which a high degree of resolution is required, Green filters are arranged in a checkerboard pattern, while in order to obtain two kinds of color signals (Red and Blue signals), Red filters and Blue filters are arranged in the checkerboard pattern.
Furthermore, as the above-mentioned color filters, there are one composed of R, G, B as shown in FIG.
2
and the other composed of the combination such as (W, G, Cy, Ye), (G, Cy, Ye) and (Mg, G, Cy, Ye), etc. wherein W represents white; Cy represents cyan; Ye represents yellow and Mg represents magenta.
In the single board sensor electronic camera as mentioned above, when employing, for example, color filters consisting of R, B, G, information on any one of R, G, B for each pixel is only obtained. Accordingly, there is a case such that the interpolation calculation on image signals is performed and each of the R, G, B data is obtained for each pixel.
For example, U.S. Pat. No. 4,642,678 discloses a structure wherein in an arrangement utilizing an R, X, B mosaic filter, when a Green signal value is obtained by the interpolation for the pixel of an s filter or a B filter, the average value of four G filter pixels adjacent to a target pixel is put as a Green signal of the target pixel. Furthermore, there is a disclosure of a structure wherein when a Red signal value and Blue signal value are obtained by interpolation, the Red signal value and the Blue signal value are linearly interpolated according to the signal value of a pixel of a Red filter and a Blue filter adjacent to the target pixel and a Green signal value interpolated at said adjacent pixel, the Green signal value interpolated at said adjacent pixel and the Green signal (original signal value for G filter pixel and interpolated value for the R and B filter pixels) at the target pixel.
Furthermore, “Digital Camera Utilizing Newly Developed Compression and Interpolation Processing” in the Proceedings of Fine Imaging Symposium (1995) of the Japan Photographic Society discloses a structure wherein upon recognizing an edge pattern, an interpolation direction is set which is pertinent for the recognized pattern, and the interpolation is performed using a pixel signal value in said interpolation direction.
Further, U.S. Pat. No. 5,373,322 discloses a structure wherein in an arrangement composed of a R, G, B mosaic filter as shown in
FIG. 2
, for the interpolation, for example, of a Green signal, the gradient of a color signal Blue and Red for an target pixel is obtained and the direction suitable for the interpolation is determined according to said gradient, and the interpolation value is obtained.
In the following, is shown examples of the interpolation calculation of G signal
34
at pixel R
34
of a Red filter shown in
FIG. 2 and a
G signal G
43
at a pixel B
43
of a B filter.
In the interpolation operation of G
34
, at first, the gradient for R
34
from R
32
, R
36
and R
14
is calculated according to the following expressions.
Hdiff=|(R
32
+R
36
)/2−R
34
|
Vdiff=|(R
14
+R
54
)/2−R
34
|
wherein Hdiff represents the gradient in the horizontal direction against R
34
and Vdiff represents the gradient in the perpendicular direction.
And, when Hdiff<Vdiff,
put G
34
=(G
33
+G
35
)/2.
When Hdiff>Vdiff
put G
34
=(G
24
+G
44
)/2.
When Hdiff=Vdiff
put G
34
=(G
24
+G
44
+G
33
+G
35
)/4.
In the same way, in the interpolation operation of G
43
, the gradient for B
43
is first calculated from B
41
, B
45
, B
23
and B
63
according to the following expressions.
Hdiff=|(B
41
+B
45
)/2−B
43
|
Vdiff=|(B
23
+B
63
)/2−B
43
|
And, when Hdiff<Vdiff,
put G
43
=(G
42
+G
44
)/2.
When Hdiff>Vdiff,
put G
43
=(G
33
+G
53
)/2.
When Hdiff=Vdiff,
put G
43
=(G
33
+G
53
+G
42
+G
44
)/4.
On the other hand, the interpolation of R signals and B signals is performed under such a structure that a linear interpolation is carried out utilizing a pixel of a G filter and signals of a R filter and a B filter adjacent closely to said pixel.
In the following, the interpolation operation expressions are shown for R
33
, R
43
and R
44
of the R signal.
R
33
=((R
32
−G
32
)+(R
34
−G
34
)/2+G
33
R
43
=((R
32
−G
32
)+(R
34
−G
34
)+(R
52
−G
52
)+(R
54
−G
54
)/4+G
43
R
44
=((R
34
−G
34
)+(R
54
−G
54
)/2+G
44
In addition, in the following, are shown operation expressions of B
33
, B
34
and B
44
of the B signal.
B
33
=((R
23
−G
23
)+(R
43
−G
43
)/2+G
33
B
34
=((R
23
−G
23
)+(R
25
−G
25
)+(B
43
−G
43
)+(R
45
−G
45
))/4+G
34
B44 =((B43 −G43)+(B45 −G45))/2 +G44
On the other hand, in the above-mentioned electronic camera, conventionally, the image data have undergone compression which are recorded in a semiconductor memory, etc. and for said compression, a method utilizing an orthogonal transformation coding has been mainly employed.
For example, in the JPEG (Joint Photographic Coding Experts) compression, the RBG signals are subjected to orthogonal transformation through DTC (Discrete Cosine Transformation) while putting 8×8 pixel as one unit, and then quantized, and Huffman coded to be compressed data. The compressed data are stored or transmitted. When the compressed data are decompressed (extended), the image is reproduced thorough the reverse process mentioned above.
As mentioned above, the orthogonal transformation coding is performed under dividing the image region into a plurality of blocks. Therefore, in the image decompression (image extension), there has been a problem such as a phenomenon wherein the joint of the blocks is not natural (hereinafter referred to as block deformation). various methods have been proposed for improving said block deformation.
For example, Japanese Patent Publication Open to Public Inspection No. 63-236088 discloses a structure wherein the orthogonal transformation coding is performed so that each block is overlapped, and Japanese Patent Publication Open to Public Inspection No. 3-166825 discloses a structure wherein a low-pass filter is applied to the portion which is judged to be a flat portion of an image. Furthermore, Japanese Patent Publication Open to Public Inspection No. 4-2273 discloses a structure wherein random noise is added to the neighbor of a block boundary or a low-pass filter is applied to that. Still further, Japanese Patent Publication Open to Public Inspection No. 6-113147 discloses a structure wherein for an image having a block deformation, a low-pass filter is applied to the portion having the deformation upon judging the presence of the deformation from the boundary.
Incidentally, in the interpolation method in the mosaic filter disclosed in the above-mentioned U.S. Pat. No. 4642678, in the Blue filter pixel, when obtaining the Red signal value by interpolation, the Green and Red si
Hung Po-Chieh
Takizawa Naruo
Frishauf, Holtz Goodman, Langer & Chick, P.C.
Garber Wendy R.
Konica Corporation
Tillery Rashawn N.
LandOfFree
Image processing method for actively edge-enhancing image... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image processing method for actively edge-enhancing image..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing method for actively edge-enhancing image... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2874732