Software development for multi-wavelength image correlation

摘要

The National Synchrotron Light Source II (NSLSII) at Brookhaven National Laboratory offers a large variety of synchrotron based imaging techniques that provide users with structural and chemical information of materials at the nanoscale. Multiple imaging techniques such as light microscopy, infrared imaging and X-ray fluorescence microscopy are commonly used to correlate information from the same sample. Correlating the important information in such images presents a large technological challenge because the various imaging techniques generate images of different sizes and spatial resolutions. To overcome this challenge, there are two goals involving software development: the first to identify a method for using fiducial markers to correlate visible light images with Xray fluorescence microscope images at the micro- to nanoscale and the second to develop software for the fusion (i.e. overlap) and correlation of these images. Numerous programs were identified to complete this project such as Matlab, Python, Photoshop, ImageJ and Fiji. After doing much research and testing a variety of these programs, it was clear that Fiji was the best and most efficient program for solving these challenges. It has the capability to align images automatically by converting the image to 8-bit gray scale, finding the maxima, i.e. darkest points, and stacking as well as corresponding these max points to align images with pin point accuracy. Moreover, it has the capability to use manually chosen fiducial point markers where the user inputs landmarks or fiducial points on the image and the program triangulates the points to align them. In addition, a user manual was written so that other synchrotron users can benefit from this methodology as well. Overall, this process will prove to be very useful in the future of the laboratory in cases where image correlation is vital. The majority of photon sciences involve correlating images and analyzing the data that come out.