|
Digital Camera Patent Abstract
A zoom tracking method for digital camera is disclosed. The method
comprises storing a position table comprising data of several zoom
position curves respect to positions of a focus lens and move the
focus lens position by looking up said position table while moving
a zoom lens position. The curves are divided into a first linear
region, a second linear region and a nonlinear region to reduce
the table size while still achieving good image quality.
Digital Camera Patent Claims
1. A zoom tracking method for digital camera, the digital camera
comprising a focus lens and a zoom lens, comprising: storing a position
table comprising data of several zoom position curves respect to
positions of said focus lens; move the focus lens position by looking
up said position table while moving said zoom lens position; wherein
said curves are divided into a first linear region, a second linear
region and a nonlinear region, and said position table comprising
a first table for storing data of said first linear region, a second
table for storing data of said second linear region and a third
table for storing data of said nonlinear region.
2. The zoom tracking method for digital camera according to claim
1, wherein said curves comprising nine zoom positions 1-9 and eight
subject positions 1 to 8 and said first linear region comprising
the first four zoom positions 1 to 4, said second linear region
comprising the middle two zoom positions 5 and 6 and said nonlinear
region comprising the last three zoom positions 7 to 9.
3. The zoom tracking method for digital camera according to claim
2, wherein in said first linear region, the eight subject positions
are separated into a lower curve group, a middle curve group and
a upper curve group.
4. The zoom tracking method for digital camera according to claim
3, wherein said the upper curve group consisting of the first two
subject positions 1 and 2, said middle curve group consisting of
the middle four subject positions 3 to 6 and said lower curve group
consisting of the last two subject positions 7 and 8.
5. The zoom tracking method for digital camera according to claim
4, wherein the value of the upper curve group is gotten from averaging
the values of the first two subject positions 1 and 2, the value
of the middle curve group is gotten from averaging the values of
the middle four subject positions 3 to 6 and the value of the lower
curve group is gotten from averaging the values of the last two
subject positions 7 and 8.
6. The zoom tracking method for digital camera according to claim
5, wherein the first table comprising data of (value of said lower
curve group+value of said middle curve+value of said upper curve
group)*(values of the four zoom positions 1 to 4).
7. The zoom tracking method for digital camera according to claim
2, wherein in said second linear region, the eight subject positions
are separated into a bottom curve, a lower curve group, a middle
curve group and a upper curve group.
8. The zoom tracking method for digital camera according to claim
7, wherein said upper curve consisting of the first subject position
1, said the lower curve group consisting of the subject positions
2 and 3, said middle curve group consisting of the subject positions
4 to 6 and said bottom curve group consisting of the subject positions
7 and 8.
9. The zoom tracking method for digital camera according to claim
8, wherein the value of the upper curve is the value of the subject
position 1, the value of said lower curve group is gotten from averaging
the values of the subject positions 2 and 3, the value of the middle
curve group is gotten from averaging the values of the subject positions
4 to 6 and the value of the bottom curve group is gotten from averaging
the values of the subject positions 7 and 8.
10. The zoom tracking method for digital camera according to claim
9, wherein the second table comprising data of (value of said bottom
curve+value of said lower curve group+value of said middle curve
group+value of said upper curve group)*(values of the two zoom positions
5 and 6).
11. The zoom tracking method for digital camera according to claim
2, wherein in said nonlinear region each value of the eight subject
positions 1 to 8 is value of a lowerbound curve+D.sub.f(d.sub.s/D.sub.s),
wherein D.sub.s is the difference between the subject positions
of an upper and a lower curves; d.sub.s is the difference between
the subject positions of an estimated and the lower curves at a
zoom start point, D.sub.f is the difference between the subject
positions of the upper bound and the lower curves at current zoom
point.
12. The zoom tracking method for digital camera according to claim
11, wherein the third table comprising data of (values of the eight
subject positions 1 to 9)*(values of the three zoom positions 7
to 9).
Digital Camera Patent Description
FIELD OF THE INVENTION
[0001] The present invention relates to a zoom tracking method
for digital camera.
BACKGROUND OF THE INVENTION
[0002] Zoom tracking adjusts a camera's focal length continuously,
to keep the in-focus state of an image during zoom operation. Zoom
tracking is important because we want the subjects always in focus
even when we are zooming in or out. In DSC (digital still camera),
we want the subjects always in focus in order to shorten the lens
moving range while auto focusing. The zoom tracking technique can
be implemented using a simple curve traced table-lookup method.
We can store several zoom position curves with respect to the in-focus
lens position, and move the focus lens position by looking up the
position table while moving the zoom lens position. FIG. 1 shows
the lens position curve. However, the table-lookup method needs
a large system memory, which is often limited in portable devices.
Another problem of zoom tracking is that the lens position curve
selection gets harder while the zoom lens moves toward the telephoto
angle. The de-focusing gradually increases as the zoom lens moves
toward the telephoto end.
[0003] One zoom tracking method in the prior art is called "Adaptive
Zoom Tracking". This algorithm uses the curve interpolation
and estimation techniques. Each curve is divided into the linear
and non-linear regions as shown in FIG. 2. Please refer to FIG.
2, in the linear region, the left and right end points are stored
in the memory and the rest focus positions are calculated from the
two points using the linear interpolation method. In the nonlinear
region, the focus position at each zoom position is obtained from
the stored curve data.
[0004] Curves between the upper and lower bound are estimated as
in Equation (1). F(k)=F.sub.1(k)-R*D.sub..eta. (1) where F(k) and
F.sub.1(k) are the focus position of the estimated and upper bound
curves at zoom position k, respectively; R is the curve estimation
factor of d.sub.f/D.sub.f, where D.sub.f is the difference between
focus position of the upper and lower bound curves at the zoom position
k, and d.sub.f is the difference of the focus position between the
upper bound and the estimation curve at the same position.
[0005] The algorithm initially traces the upper bound curve since
the difference between the focus positions of each curve is very
small in the linear region; in non-linear region, the curve estimation
factor is calculated. Finally, zoom curve interpolation and estimation
are performed using the curve estimation method.
[0006] However, the adaptive zoom tracking mentioned above only
traces the upper bound curve, where the loss of steps will get larger
while the curve is closer to the lower bound.
[0007] In views of the above-described disadvantages resulted from
the prior art, the applicant keeps on carving unflaggingly to develop
a wireless mouse according to the present invention through wholehearted
experience and research.
SUMMARY OF THE INVENTION
[0008] It is an object of the present invention to provide a zoom
tracking method for digital camera which can reduce size of data
while still deliver good image quality.
[0009] This invention providing a zoom tracking method for digital
camera, the digital camera comprising a focus lens and a zoom lens,
comprising: storing a position table comprising data of several
zoom position curves respect to positions of said focus lens, and
move the focus lens position by looking up said position table while
moving said zoom lens position; wherein said curves are divided
into a first linear region, a second linear region and a nonlinear
region, and said position table comprising a first table for storing
data of said first linear region, a second table for storing data
of said second linear region and a third table for storing data
of said nonlinear region.
[0010] In a preferred embodiment, the curves comprising nine zoom
positions 1-9 and eight subject positions 1 to 8 and said first
linear region comprising the first four zoom positions 1 to 4, said
second linear region comprising the middle two zoom positions 5
and 6 and said nonlinear region comprising the last three zoom positions
7 to 9. In said first linear region, the eight subject positions
are separated into a lower curve group, a middle curve group and
a upper curve group. The lower curve group consisting of the first
two subject positions 1 and 2, the middle curve group consisting
of the middle four subject positions 3 to 6 and the upper curve
group consisting of the last two subject positions 7 and 8.
[0011] In a preferred embodiment, the value of the lower curve
group is gotten from averaging the values of the first two subject
positions 1 and 2, the value of the middle curve group is gotten
from averaging the values of the middle four subject positions 3
to 6 and the value of the upper curve group is gotten from averaging
the values of the last two subject positions 7 and 8.
[0012] In a preferred embodiment, the zoom tracking method for
digital camera according to claim 5, wherein the first table comprising
data of (value of said lower curve group+value of said middle curve+value
of said upper curve group)*(values of the four zoom positions 1
to 4).
[0013] In a preferred embodiment, in said second linear region,
the eight subject positions are separated into a bottom curve, a
lower curve group, a middle curve group and a upper curve group.
The bottom curve consisting of the first subject position 1, said
the lower curve group consisting of the subject positions 2 and
3, said middle curve group consisting of the subject positions 4
to 6 and said upper curve group consisting of the subject positions
7 and 8.
[0014] In a preferred embodiment, the value of the bottom curve
is the value of the subject position 1, the value of said lower
curve group is gotten from averaging the values of the subject positions
2 and 3, the value of the middle curve group is gotten from averaging
the values of the subject positions 4 to 6 and the value of the
upper curve group is gotten from averaging the values of the subject
positions 7 and 8.
[0015] In a preferred embodiment, the second table comprising data
of (value of said bottom curve+value of said lower curve group+value
of said middle curve group+value of said upper curve group)*(values
of the two zoom positions 5 and 6).
[0016] In a preferred embodiment, in said nonlinear region each
value of the eight subject positions 1 to 8 is value of a lowerbound
curve+D.sub.f(d.sub.s/D.sub.s), wherein D.sub.s is the difference
between the subject positions of an upper and a lower curves; d.sub.s
is the difference between the subject positions of an estimated
and the lower curves at a zoom start point, D.sub.f is the difference
between the subject positions of the upper bound and the lower curves
at current zoom point.
[0017] In a preferred embodiment, the third table comprising data
of (values of the eight subject positions 1 to 9)*(values of the
three zoom positions 7 to 9).
[0018] The above objects and advantages of the present invention
will become more readily apparent to those ordinarily skilled in
the art after reviewing the following detailed description and accompanying
drawings, in which:
BRIEF DESCRIPTION OF THE DRAWINGS
[0019] FIG. 1 shows an example of lens position curves.
[0020] FIG. 2 shows linear and nonlinear regions of lens position
curves of the Adaptive Zoom Tracking method of prior art.
[0021] FIG. 3 shows the lens position curves with nine zoom positions.
[0022] FIG. 4 shows the lens position curves divided into linear
region 1, linear region 2 and nonlinear region of this invention.
[0023] FIG. 5(A)-(I) shows experiment results of this invention
and the prior art method.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
[0024] Please refer to FIG. 3, it shows an example of lens position
curves of this invention. We first cut the zoom step range into
nine zoom positions, zoom 1 to zoom 9. For every zoom positions,
we measure the focus step of eight different subject distances to
construct zoom table. Therefore, we have 9*8=72 data shown in FIG.
3.
[0025] Our goal is to reduce the data while still performing good
zoom tracking image quality. We divide the data curve into Linear
Region 1, Linear Region 2, and Nonlinear Region, as shown in FIG.
4, which means that we separate the data into three tables. However,
we do not keep all the data; we only reserve the complete data in
the Nonlinear Region.
[0026] In Linear Region 1, we extract the first four zoom positions'
focus step from the original data. We will have 4*8=32 data, but
we do not store all of them. Instead, we separate eight subject
positions into three groups, the lower curve, the middle curve,
and the upper curve. The lower curve consists of the first two subject
positions, which is 0.4 m and 0.5 m, for nine zoom positions, and
the values are gotten from averaging the values of two close subject
positions, and both subject positions are set to the same value.
The middle curve contains the middle four subject positions, 0.6
m, 0.8 m, 1 m, and 1.5 m. The value is set to the average value.
The upper curve contains the last two subject positions, 2.5 m and
infinity, and set to the average value. In this way, the original
32 data are reduced to 3 (lower, middle, and upper curve)*4 (zoom
positions)=12 data, where 20 data are saved. The three curves are
shown in the Linear Region 1 in FIG. 4.
[0027] Linear Region 2 uses the same concept to divide the data
for the middle two zoom positions, which are 2*8=16 data. In this
region, data are more dispersed. Therefore, we add one curve, bottom
curve, for accuracy. Bottom curve conserves the original data of
0.4 m. Lower curve sets the value by averaging 0.5 m and 0.6 m's
focus step. Middle curve contains 0.8 m, 1.0 m, and 1.5 m. Upper
curve uses the average value of 2.5 m and infinity. Therefore, only
4 (bottom curve value+lower curve group value+middle curve group
value+upper curve group value)*2 (zoom positions)=8 data will be
stored in this region, which are 8 data saved. The four curves are
shown in the Linear Region 2 in FIG. 4.
[0028] Once the data of Linear Region I and Linear Region 2 are
stored, we can find our zoom tracking focus steps by directly looking
up the table. The estimated curve's focus step is obtained by comparing
the previous focus step to tables, finding its related focus step
index, and getting next zoom position's focus step directly from
the table.
[0029] In the nonlinear region, all curves are almost separated,
so all data of the last three zoom positions are stored. The estimated
curve's focus step is set by curve interpolation method in Equation
(2). Estimated = LowerBound + D f .times. d s D s ( 2 )
[0030] where Estimated represent the expected focus position we
are looking for; LowerBound means the minimum focus position during
the experiment; D.sub.s is the difference between the focus positions
of the upper and the lower curves; d.sub.s is the difference between
the focus positions of the estimated and the lower curves at the
zoom start point; D.sub.f is the difference between the focus positions
of the upper bound and the lower curves at current zoom point.
[0031] Combining the data of linear region 1, linear region 2 and
nonlinear region, the stored data are 12 (first linear region)+8
(second linear region)+8*3 (nonlinear region)=44 data and the original
data are 72 data. Therefore, the data is reduced by 44/72=61%, almost
1/3 data are saved.
[0032] FIG. 5A-5I shows the experimental result of the reduced
zoom tracking method (called RZT) of this invention and the adaptive
zoom tracking method in the prior art which just traces the upper
bound in the linear region (called UZT). FIG. 5 shows the results
in visual comparison, and focus step error comparing with the focus
step found by global search (GS). FIG. 5 explains the images from
wide-angle to tele-photo.
[0033] FIG. 5(A) is the original, non-zooming image, and its focus
step by GS is 26. We use the global search to ensure the best focus
position. We start from this image, and begin zoom tracking. These
images are down-sampled one-fourths. Compare 5(B1) and 5(B2), which
is zoom position 1 in RZT and UZT, respectively. As we can see,
5(B1) is sharper than 5(B2). The focus step derived by RZT is 34,
which is close to 32; the UZT focus step is 44, which means it has
12 steps error. The first five zoom steps are in linear region,
and the last three are in non-linear region. RZT in FIGS. 5(B1)
to 5(F1) are all sharper than UZT in FIGS. 5(B2) to 5(F2), and the
focus steps of RZT are all closer to GS focus steps, while UZT all
have large errors. In the non-linear region, compare 5(G1) and 5(G2),
we can discover that 5(G1) still has the focus step close to GS,
with 5(G2) still has a large error. However, when compared with
5(H1) and 5(H2), 5(H1) has a large step error in this scenario,
while the focus step of RZT provides a larger error. In the final
two figures, 5(I1) and 5(I2), RZT is still close to the GS focus
steps.
[0034] To summarize, in the lens moving direction from wide-angle
to tele-photo, our RZT method is shown to be more accurate than
UZT while still shrinking the data size.
[0035] While the invention has been described in terms of what
is presently considered to be the most practical and preferred embodiments,
it is to be understood that the invention needs not be limited to
the disclosed embodiment. On the contrary, it is intended to cover
various modifications and similar arrangements included within the
spirit and scope of the appended claims which are to be accorded
with the broadest interpretation so as to encompass all such modifications
and similar structures.
|