Image analysis – Image enhancement or restoration – Focus measuring or adjusting
Reexamination Certificate
1998-12-16
2002-01-22
Lee, Thomas D. (Department: 2724)
Image analysis
Image enhancement or restoration
Focus measuring or adjusting
C382S263000
Reexamination Certificate
active
06341180
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to systems (procedures and/or devices) for autofocusing according to the content of an image and, more particularly, to an autofocusing system for microscopes and similar instruments which is insensitive to noise.
BACKGROUND OF THE INVENTION
Microscopes and other optical instruments need to focus on an object of interest. Preferably, only one focusing step is performed for each object and, once focused, the object will remain in focus. Ideally, for an instrument which scans an object, the object remains in focus while scanning is performed. However, the object analyzed by optical instruments is typically not flat. In addition, not all objects of interest are situated at the same depth with respect to the objective lens. Accordingly, some refocusing is required for such scanning instruments.
With scanning instruments, once a crude focus position has been found, and if the scanning of the object is done in small steps, fine focus adjustments can be sufficient to keep the object in focus so long as the adjustments are performed after each step (movement). For a microscope's objective lens with a numerical aperture of 1.3, a fine focusing adjustment with an accuracy of about ±0.1 micrometers can be required after each step. The procedures used for crude and fine focusing can be the same, although the range of the search and the search step length are usually different. However, if the fine focusing procedure is computationally too intense, another, less intense, more approximate, procedure may be used for crude focusing.
A typical autofocusing procedure system primarily: (1) estimates what the focus (distance and/or angle of the objective lens relative to the object) should be; and (2) mechanically adjusts the distance and/or angle by the focusing mechanism using a predetermined search strategy which is implemented using a controller. Some autofocusing systems use image content information to estimate the focus. Such systems are often called ‘passive autofocusing systems’. In such an image content autofocusing system, the image detected by the microscope is provided to a subsystem which estimates the focus.
One benefit of image content autofocusing systems is that the calculations can be performed in software, that is, extra hardware is not needed, thereby reducing the cost of the system. Also, with an image content autofocusing system, any discrepancy (drift) between the estimated focus calculated by the focusing subsystem and the ability of the focusing mechanism to implement the estimated focus is detectable and can be eliminated. Typical applications for image content autofocusing systems include inspecting chips in the semiconductor industry and all kinds of medical microscopy. In addition, if an image content autofocusing system does not need to have highly repeatable focusing mechanics, the focusing mechanism cost can be considerably reduced.
The performance of an image content autofocusing system is primarily limited by the desired search speed for the instrument and the noise-sensitivity of the focus estimation procedure. The search speed depends, among other parameters, on: (1) the rate at which images of the object are acquired (video frame rate); (2) the speed at which the focusing mechanism can respond (mechanical time constants); (3) the search range, once focused; (4) the search step length; (5) the amount of time required for image processing; and (6) the computing power available. In addition, interference from noise must be taken into account by the focus estimation procedure or the optimal focus can not be achieved. In other words, if the focus estimation subsystem is noise-sensitive, the focus estimate will not correspond to the optimal (maximum) desired focus position.
Two typical image content autofocusing procedures which are noise-sensitive maximize the statistical variance of the content of the raw image or the statistical variance of a high-pass filtered version of the image content. Despite the noise-sensitivity, these statistical variances can be used for finding a crude focus position.
To correct for the noise-sensitivity, some conventional procedures measure the statistical variance at multiple foci in the search range, and then using curve-fitting techniques better estimate the optimal (maximum) focus for the objective lens. See, for example, U.S. Pat. No. 5,790,710.
However, such post-processing (a posteriori) curve-fitting techniques have several drawbacks: (1) the mechanical step lengths of the focusing mechanism must be precise for the curve-fitting technique to be accurate; (2) the optimal focus is not known until after all the images used by the procedure have been taken; and (3) the optimal focus may be near the edge of the curve, which can adversely affect the curve-fitting.
Correcting for these drawbacks is usually costly in terms of time or expense because, for example: (1) a more expensive focusing mechanism must be used for precise positioning; (2) more storage may be required to store all the images until the image corresponding to the optimal focus can be determined; and (3) the entire procedure may have to performed again, starting at a different focus estimate, to ensure that the optimal focus is not near the edge of the curve. Although one solution to having more storage takes advantage of a precise positioning focusing mechanism required for other purposes, going back to grab an image at the optimal focus after the post-processing can have problems caused by hysteresis in the control of the focusing mechanism and can take longer to locate the optimal focus.
SUMMARY OF THE INVENTION
This invention provides solutions to the problems encountered with conventional autofocusing systems by employing a noise-insensitive filter which operates on image signal data from “grabbed” images of the already existing microscope camera.
One goal of the invention is to provide an autofocusing system which uses a low-cost focusing mechanism by having a filter calculator employ a filter capable of achieving the optimal focus using the content of the image (focus function) without the need for any curve-fitting technique to maximize the measured amount of focus (focus metric) from an energy calculator.
Another goal of the invention is to minimize the amount of time required to obtain a sharp focus.
One object of the invention is to provide a device for determining a focus between an image sensor and an object having: (1) an image sensor for receiving an image of the object and for generating an image signal from the image, the image signal having an image component and a noise component; (2) a filter calculator for receiving the image signal and for generating a filtered image signal such that the noise component of the image signal has been reduced, the noise component being reduced by increasing the energy contributions of parts of the image signal which contribute a relatively larger proportion to the image component than the noise component and by decreasing the energy contributions of other parts of the image signal which contribute a relatively larger portion to the noise component than to the image component; (3) an energy calculator for receiving the filtered image signal and determining an energy level of the filtered image signal; (4) a controller for receiving the energy level and for generating a position control signal in accordance with the energy level; and (5) a positioning mechanism for receiving the control signal and performing a focus adjustment based on the control signal. This filter calculator can also have a linear convoluter; and a filter selected from the group consisting of: a Wiener filter, a modified second order difference filter, and a bandpass filter. The filter can be a one-dimensional array. Also, the filter calculator can be limited to only operate on: (1) the image signal in accordance with a single direction in the image; and/or (2) portions of the image signal corresponding to selected sections of the image.
In addition, the device can include a filter
Pettersson Magnus
Rosenqvist Anders
Brinich Stephen
Cellavision AB
Cooper & Dunham LLP
Lee Thomas D.
LandOfFree
Image content autofocus for microscopy using a... 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 content autofocus for microscopy using a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image content autofocus for microscopy using a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2865424