The overall goal of this procedure is to synthetically increase the lens aperture of a scanning imaging platform. This is accomplished by first capturing three low resolution, differently defocused images of a target while moving along the optical axis of the system. The second step is to shift the entire imaging system perpendicularly, and then take three defocused images of the target.
Next, shift the imaging system to the other side of the optical axis and capture three more defocused images of the target. The final step is to numerically retrieve the optical phases in order to determine the optical fields and then properly combine them to achieve a super resolved image. Ultimately, a synthetically increased lens aperture is generated along the direction of movement yielding higher imaging resolution.
The main advantage of the proposed technique over other time multiplexing super resolving approaches is that our technique is passive, and therefore it does not require projection of encoding patterns that are later unused to obtain the super resolved image, This experiment is normally performed in relative darkness. However, some of the video is filmed with the light on In order to better visualize the protocol, begin set up with the rough alignment of the laser beam expander lens and camera on the same optical axis. Mount both the lens and camera on a translation stage to allow for subtle movements perpendicular to the optical axis.
Additionally, mount the camera on a translation stage for small motion parallel to the optical axis. Turn on the laser and use an aperture iris to make sure the light passes through the center of the lens. Next, turn on the camera and use the linear Z stage to check the alignment of the laser beam.
When it is aligned, defocusing, the camera will only cause the spot to change its size, but will not cause lateral shift of the spot. After alignment has been completed, insert a US Air Force test target in front of the beam expander. Place the target so that light passing through it passes through the center of the lens.
Use the linear Z stage in order to focus the target. This initial xz position of the camera will serve as the anchor point. Once focus is achieved, insert the 0.1 inch square aperture and capture the first image of the target.
Now adjust the linear Z stage. Use it to shift the camera away from the target 0.2 inches. Capture a second image of the target, move the camera another 0.2 inches away.
Take a third image. These three images will be referred to as the B series. Return the camera to its original anchor position before proceeding.
After returning to the anchor position, begin to use the linear X stage shift. The entire imaging system laterally a distance of positive 0.1 inches. The imaging system is now off center of the laser beam.
Capture an image of the target from this position. Adjust the Z stage to move the camera away from the target 0.2 inches. Capture an image, then move it back another 0.2 inches.
Take a third image of the target. These three images will be referred to as the A series. Return the camera to the anchor position.
Starting at the anchor position, shift the camera negative 0.1 inches. Capture three more images at the same Z positions as the other series. These images will be the C series.
Since the camera captures only the field intensity, the optical phase information is lost. In order to recover it and find the optical field, make use of the numerical three planes method. Once the optical field of each image series is found, use the phenal free space integral to back propagate the optical field of the B series to the lens playing for the A series.
Make certain the field is shifted to reflect its position with respect to the optical axis. Free space propagate its optical field to the lens plane. Repeat the same steps for the C series below the optical axis.
Sum the three fields to combine them and synthetically increase the aperture size. Finally, free space propagate the resulting field to the image plane. The target used in the experiment was the negative 1951 USAF test target shown here in a high resolution image.
Compare this to the low resolution image taken at the anchor position on the optical axis. None of the resolution bars are visible in the super resolved image. The vertical bars are visible up to the third element on the right.
Because the aperture was only increased in the horizontal direction X, there is no improvement in resolution of the horizontal bars. After watching this video, you should have a good understanding of how to a passive super resolution system that synthetically increases the lens aperture by using the motion of the imaging platform and numerical computation. Although the demonstration you saw was on an optical bench, the proposed concept is feasible for real airborne imaging systems.
