next up previous
Next: mpiv_filter.m Up: Reference Manual of mpiv Previous: Reference Manual of mpiv

mpiv.m

mpiv.m is the main routine of mpiv. mpiv.m is an external function (.m file) in MATLAB with its format as following:

function [xi,yi,iu,iv]=mpiv( im1, im2, nx_window, ny_window, ...
                             overlap_x, overlap_y, ...
                             iu_max, iv_max, dt, ...
                             piv_type, i_recur, i_plot).

The input augments of mpiv.m are:

im1 and im2    : input images (double precision)
{nx,ny}_window : processing window sizes in x and y, 
                 -> 16, 32, or 64 are recommended (unit: pixel)
                 64 is good to start with in recursive PIV.
overlap_{x,y}  : fraction of overlapping windows in x and y	
                 -> range [0.0 to 0.9] (no overlap to 90% overlap)
                 0.0 or 0.5 is typically used.
{iu,iv}_max    : maximum allowed velocities in x and y (in pixel)
                 If entering 0, they become 1/3 of the window size 
                 -> these two parameter also provide the search area 
                    if `mqd' is chosen
                    (A large value will increased CPU time dramatically).
dt             : time separation between im1 and im2 (unit: second)
                 use `1' to give output in pixel
piv_type       : = `cor': cross-correlation algorithm
                 = `mqd': MQD algorithm
i_recur        : number of recurrence to enhance vector resolution/accuracy
                 = 0 -> no recurrence
                 = 1 -> one recurrence with the same window size
                 = 2 -> reduce window size by one half
                 each increment (after 1) reduces the window size by one half 
i_plot         : = 1 -> plot the calculated results
                 otherwise -> no plot

The output from mpiv.m is as follows:

xi, yi         : location of vectors in x and y (in pixel)
iu, iv         : output velocity vectors in x and y (in pixel if dt = 1.0)
Please mind the arrangement/orientation of the output arrays and matrices in the programs. It normally requires matrix transpose for plotting.




2006-12-08