Scott Hillberry

LIHSP ( under Federica Bianco )

Lucky Imaging with High-Speed Photometry (LIHSP) can best be described as the combination of quick exposures and image quality assessment. Due to atmospheric aberrations, it is often more desirable to take numerous quick exposures (~ milliseconds) and then stack them to obtain an effectively long exposure. Lucky Imaging itself involves using a small percentage of the total images to create the final stack. Using only the best images (usually defined by their Strehl ratio or a similar quality) increases image resolution.

Using this technique, we are able to obtain much greater resolution than a normal exposure would yield. This method is used to analyze closely-spaced objects, such as those in a binary system or microlensing field.

Phase Correlation

When lucky imaging is used on localized objects (e.g. stars), it is not inaccurate to treat the field as a point source represented by the brightest pixel in the field. In this case, stacking the images is a simple matter of stacking the brightest pixels.

However, this model begins to fail when used on more diffuse or distended objects such as fields with nebulae and other emissions. In these cases, the brightest pixel from image to image is not as fixed. Instead, it is more effective to convert the images to frequency-space (using Fourier transforms) and match the points of greatest correlation. Simply put, the phase correlation method (PCM) looks for repeating patterns between two images.

Provided by Fredrik Orderud in the Wikimedia Commons

Wavelet Sharpness Assessor

The relative sharpness of an image is a very important quality for both assessing the success of image processing and comparing one image to another. This is especially important when defining the "best" images for lucky imaging. Because the Strehl ratio concerns the point spread function (psf) of a source, it naturally cannot describe a distended source well. Hence, a different method is necessary to compare image quality.

By performing a Wavelet transform on an image, the edges are enhanced and can be assessed by examining so-called Lipschitz exponents¹. These exponents are a measure of the change in Wavelet transform modulus at a given point in relation to a change in dilation scale of the transform. Sharper images have lesser exponents in relation to less sharp images.