Online image analysis

Aim : Write a basic image analysis package, with front end Qt

  • Measure positions and brightness of stars
  • Mac development via ports : py27-matplotlib (+cairo +tkinter +latex +pyside +qt4), py27-scipy, py27-ipython, pip install iminuit (py27-pip),
  • Windows :
-- WillParker - 16 Jun 2014

Log:

  • 16 June: Implemeneted flood fill algorithm utilised to determine numbers of stars present in a given data file.
  • 17 June: Testing work and improving performance where possible. Testing methods to automatically determine a threshold for a given data file. If time allows, will go on to determining a distance between identified stars.
-- WillParker - 17 Jun 2014

  • 19 June: Script written that identifies clusters of points from a given .fit data file associated with an intensity above a given threshold and masks them on a plot. Analysis script written that returns the centre (using the weighted mean), widths and heights of said clusters.
cluster-centre.png
  • 20 June Investigating (and hopefully implementation) RGB alpha overlay so that the masking of the clusters can be made transparent. Extension of the analysis script to include root mean square and full width half height maximum (FWHM).
-- WillParker - 20 Jun 2014

  • 23rd June. Modified Cluster_Finder class so that the overlaying plot transparency can be adjusted (useful when implementing Qt interface later). Cluster_Analysis class wrapped up and useful stats about identified clusters produced. Git adopted as a version control system for the rest of the project, first pushes to master made (hopefully breaking nothing), repository can be found on bitbucket here. Read up on photometry aperture in order to plan rest of work through the week. Still need to implement a 2d Gaussian fit of an identified cluster, planning to use curve fit for this but unsure how to proceed currently as not quite sure what I am fitting. Assuming it is the intensities f(z) for each coordinate point z = (x,y).
-- WillParker - 23 Jun 2014

  • 27 June. Created Photmetry_Aperture analysis which successfully calculates and plots relevant aperture measurements of identified clusters of stars. Full width at half height maximum method has been implemented whilst other statistical methods have corrected (RMS, Std dev, etc).
clusters.pngcluster_aperture.png

Here the red circles identify the varying photometry aperture. Various cosmetic changes were undertaken such as changing the colourmaps for the plots and setting transparency (rgb alpha) values. Came up with an idea to create a Class_Tester class which enables easy testing of the developing codebase helping to ensure that any broken code does not get pushed to the master branch. I came up with a method to count the number of pixels within the bounding circles marked on the plots above. Here is a sample of the statistical output of the program to a terminal window thus far:

cluster_output.png

-- WillParker - 27 Jun 2014

Topic attachments
I Attachment Action Size Date Who Comment
PNGpng all-clusters.png manage 107.7 K 20 Jun 2014 - 09:35 WillParker  
PNGpng cluster-centre.png manage 90.5 K 20 Jun 2014 - 09:40 WillParker  
PNGpng cluster_aperture.png manage 49.8 K 27 Jun 2014 - 17:09 WillParker  
PNGpng cluster_output.png manage 26.1 K 27 Jun 2014 - 17:18 WillParker  
PNGpng clusters.png manage 75.6 K 27 Jun 2014 - 17:00 WillParker  

Physics WebpagesRHUL WebpagesCampus Connect • Royal Holloway, University of London, Egham, Surrey TW20 0EX; Tel/Fax +44 (0)1784 434455/437520

Topic revision: r6 - 27 Jun 2014 - WillParker

 
This site is powered by the TWiki collaboration platformCopyright © 2008-2018 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding RHUL Physics Department TWiki? Send feedback