Automatic focus for image sensors

• US 7,589,781 B2
• Filed: 01/26/2005
• Issued: 09/15/2009
• Est. Priority Date: 01/26/2005
• Status: Active Grant

• Current Assignee: Omnivision Technologies Incorporated (Will Semiconductor Co., Ltd. Shanghai)
• Original Assignee: Omnivision Technologies Incorporated (Will Semiconductor Co., Ltd. Shanghai)
• Inventors: Huang, Xiaolin
• Primary Examiner(s): Chan; Jason
• Assistant Examiner(s): OSINSKI, MICHAEL S
• Application Number:
  US11/044,137
• Publication Number:
  US 20060164934A1
• Time in Patent Office: 1,693 Days
• Field of Search: 348345-357, 348/208.12, 348/207.99, 348/208.16, 348/222.1, 369/44.11
• US Class Current: 348/355
• CPC Class Codes: H04N 23/673 based on contrast or high frequency components of image signals, e.g. hill climbing method

Abstract

The embodiments of the invention relate to automatic focusing methods and apparatus suitable for use in digital imaging devices. The proposed methods and apparatus evaluate the focus condition of the optical system by evaluating the digitized image, without requiring additional hardware besides the digital image-capture signal flow path. A method for computing the appropriate direction and magnitude of optical system adjustment, based on the evaluated focus condition, is provided. The computed quantities are communicated to the lens driving apparatus to reposition and adjust the optical system and to incrementally obtain a better focus. Upon reaching a desired focus quality, the device is prevented from controlling the lens driving apparatus.

Claims

1. An automatic focusing apparatus as a part of a digital imaging device, the apparatus comprising:

an optical system capable of forming an image of an object at a focal plane of the optical system;

a two dimensional away of light sensors for generating pixel data representing the image;

a windowing facility cooperating with a plurality of filter taps, the windowing facility being configured for;

extracting the pixel data of at least one region of the image using a first finite impulse response (FIR) filter to derive first spectral information and a second FIR filter to derive second spectral information, the first and the second FIR filters being further configured for delaying the pixel data, multiplying the taps by the delayed pixel data, and accumulating results of multiplying the filter taps by the delayed pixel data to generate a first digitally filtered sequence of pixel data and a second digitally filtered sequence of spectral information, respectively;

computing at least one value corresponding to a focus condition of the region based on the region'"'"'s extracted pixel data; and

computing at least one value corresponding to a scene activity of the region based on the region'"'"'s extracted pixel data;

a focus condition change detection facility for computing an out-of-focus indicator value based on the scene activity value;

the windowing facility further comprising;

at least one window processing resource for each region of the image, wherein each window processing resource implements a digital filter to process the said extracted pixel sequence to produce a digitally filtered sequence; and

a plurality of accumulator units to accumulate each of the said digitally filtered sequences by adding the pixel data in the sequence to a total value representing the sum of all previous pixel data in the sequence, wherein a final accumulated result is a total value computed by the accumulator unit after adding the last pixel data in the digitally filtered sequence;

the windowing facility being further configured for;

computing arithmetic weighted accumulated sum by weighting and summing the final accumulated results of window processing resources;

wherein

arithmetic weights are partitioned into at least one curve group, and where each curve group is defined by containing exactly one final accumulated result from a window processing resource and associating each final accumulated result with a curve group;

the focus condition change detection facility is configured for computing two sets of intermediate results for an out-of-focus indicator value using results from at least two of the curve groups by operating on the weighted accumulated sums of each curve group, wherein;

the first set of intermediate results is computed by taking the difference between two predetermined weighted accumulated sums from the same curve group;

the second set of intermediate results is computed by dividing two predetermined weighted accumulated sums from the same curve group;

the out-of-focus indicator is computed by a threshold detection facility that performs logical operations on four sets of values, wherein the four sets of values comprise two sets of intermediate results from a current image and two sets of intermediate results from a previous image; and

a tracking state-machine for computing a direction value and a magnitude value for adjusting and repositioning the optical system from a current position, based on the computed focus condition value and the out-of-focus indicator value.

2. The apparatus of claim 1, wherein the tracking state-machine'"'"'s computations of the direction value and the magnitude value comprise:

storing the focus condition values of previous positions, where;

in a first mode, the focus condition values of the two previous consecutive positions are stored;

in a second mode, the focus condition values of a position having the best focus quality is stored and the other values are discarded; and

in a third mode, all stored focus condition values are discarded;

comparing the current focus condition values to the stored focus condition values and determining a desired position for adjusting the optical system, wherein the direction value and the magnitude value are computed based on the stored focus condition values and the previous direction and magnitude values;

stopping the focus adjustment when a minimum magnitude quantity is reached and the focus condition is determined to be more optimal than any stored previous focus condition; and

restarting the focus adjustment when the out-of-focus indicator value signifies an out-of-focus situation.

3. The apparatus of claim 1, wherein an image region is specified by providing the windowing facility its location and size, and wherein the regions may overlap.

4. The apparatus of claim 1, wherein the windowing facility comprises at least one main window processing resource and one secondary window processing resource, and wherein:

the main window processing resource is dedicated to video auto-focusing applications;

the secondary window processing resource is dedicated to image auto-focusing applications; and

subsequent processing of the main window processing resource results and the subsequent processing of the secondary window processing resource results are performed independently.

5. The apparatus of claim 4, further comprising a mode to select between the said main window processing resource and the said secondary window processing resource as an active processing resource in the said windowing facility.

6. The apparatus of claim 1, wherein a set of digital filter coefficients used in each said digital FIR filter is selected from one or more predetermined sets of digital filter coefficients.

7. The apparatus of claim 1, wherein arithmetic weights can be updated for each new image.

8. The apparatus of claim 1, wherein the mentioned processes are programmed into at least one integrated circuit of the digital imaging device.

9. The apparatus of claim 1, wherein window processing resources operate on one pixel sequence from a single readout of the image.