1 DG_SMOOTH DG_SMOOTH Smoothes an image using the Conjugate Gradient Algorithm. (Author: Didier GIRARD, Groupe d'Astrophysique) The smoothed image is the equilibrium state of a thin flexible plate constrained to pass near each height datum by a spring attached between it and the plate. The smoothing is controled by the "Smoothing Parameter" P : P = ( plate stiffness) / ( springs stiffness) A low value for P (0.001) means high fidelity to the original data and should be used for high signal to noise ratios. If the data are noisy, use higher values of P (0.1 to 10). The algorithm is iterative and needs a work space equal to 3 input maps. It usually converges in 10 iterations, and can be restarted if you save the work files. Timing is of the order of 2 seconds of microVAX II CPU time for a 128 by 128 map. It takes into account blanked pixels properly. 2 RESTART$ TASK\LOGICAL "Restart smoothing from previous step" RESTART$ This is a logical flag to indicate restart from a previous run of the program. 2 P$ TASK\REAL "Smoothing parameter" P$ This is the value of the smoothing parameters. It should remain between 0.001 and 10. Higher values give stronger smoothing. 2 NITER$ TASK\INTEGER "Number of smoothing iterations" NITER$ This is the number of smoothing iterations. Ten iterations are usually enough to reach convergence. 2 Z_NAME$ TASK\FILE "Input file name" Z_NAME$ This is the name of the input file. 2 Y_NAME$ TASK\FILE "Output file name" Y_NAME$ This is the name of the internal work file. It is not deleted automatically, and it is needed if you want to restart from a previous number of iterations. 2 X_NAME$ TASK\FILE "Work file name" X_NAME$ This is the name of the output smoothed image. 2 GUESS$ TASK\REAL "Initial value for blanked pixels" GUESS$ This is an initial guess for the value attributed to blanked pixels. Try to find something reasonable. Convergence can be delayed if you choose a crazy value. 1 END