XMM-Newton Science Analysis System
omdetect (omdetect-5.14.2) [???]
Meta Index / Home Page / Algorithm / The detection algorithms
omdetect uses several validation functions, and the discussion that follows
refers to the one that is used the most frequently.
This function works as follows:
- Initialise xc to real(i0) + 0.5 , where i0 is the x-index of the current pixel being checked.
- Initialise yc to real(j0) + 0.5 , where j0 is the y-index of the current pixel being checked.
- Initialise the loop counters l to 1, m to 1, i to the number of input circles and j to 1, where
m and n range from 1 to 5, i from 1 to the number of input circles and j from 1 to the number
of points round each circle
- Initialise xCentre to xc + xOffset(l), where xOffset has values from -0.5 to 0.5 in steps of 0.25
- Initialise yCentre to yc + yOffset(m), where yOffset has values from -0.5 to 0.5 in steps of 0.25
- Initialise xPos to xCentre + rcosTheta(j, i) and yPos to yCentre + rsinTheta(j, i), where
rcosTheta and rsinTheta store pre-computed values of the quantities r x cos(theta) and r x sin(theta),
for a given radius (r) of the circle and values of theta from 0 to 360.
- Compute the value of the image at the point xPos and yPos by bi-linear interrpolation and store
the value in the array fcircle.
- When all the points round the circle have been computed, compute the mean and variance
of these values, and store these values in mean(i) and sigma(i).
- If i equals the number of circles, decrement i by 1 and repeat the process for the next circle.
- Otherwise, compare the difference between the mean(i) and the mean(i+1) and if this
difference is greater than or equal to the quantity nstd
sqrt(sigma(i) + sigma(i + 1) (where nstd is an input
parameter), conclude that there is a valid source centred on xCentre, yCentre, set the function to true
and return.
- Otherwise try another x/y offset.
- If all the offsets have been looped through, set the function to false and return.
XMM-Newton SOC/SSC -- 2008-10-22