Abstract
Purpose :
To achieve an optimal correction in Adaptive Optics (AO) systems, the size of the imaged region should be smaller than the isoplanatic angle (~2 deg for human eye). Most AO systems align videos after the imaging session. In patients with poor fixation, eye movement can make it difficult to align most frames of the videos.
Methods :
The Indiana AOSLO system images the retina using a wavelength of 820 nm capturing videos of ~2x2 degrees (100 frames, 30Hz). We designed a system to capture simultaneously images with a larger angle centered on the area imaged by the AOSLO. The system was designed in Zemax, to optically match the main system. The subsystem uses a cylindrical lens to project a 6 deg horizontal line on the retina with a wavelength of 769 nm. It was coupled to the AOSLO using a dichroic beamsplitter. Using the vertical scanner of the AOSLO, the line is moved across the retina and an image is recorded sequentially with a line scan camera (Basler Industrial Cameras, spl2048-70km). We operated the two systems simultaneously and recorded the images of a subject capturing videos (~2x2 deg) with the AOSLO of a region of the photoreceptor mosaic and, by changing the size of the deflections of the vertical scanner, videos with the subsystem over an area of 5.5x3.3 deg. To test the method, custom software was developed to process the images from the subsystem and create an averaged image that could be used as the template to which the individual frames of the small field AOSLO can be aligned.
Results :
In the subsystem design, a diffraction limited image quality was achieved within the imaging region (RMS wavefront errors are 0.01 λ for 0 degree and 0.1 λ for ±3 degrees). In practice the wide field image was able to image cones. The 5.5x3.3 degrees image recorded with the subsystem were aligned and the average image was used as a template to align some frames of the video captured with the AOSLO (2x2 degrees). The figure shows the averaged image used as template and some of the AOSLO video registered to the template. Frames from different videos imaging nearby retina could be aligned to the same template.
Conclusions :
A line scan imaging system can be used to simultaneously capture a larger area of the retina. This image, though it has a lower resolution, can be used to improve the alignment efficiency.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.