The FILTER Sub-Menu

The FILTER sub-menu contains commands which allow filtering and smoothing of data. (Other commands exist in the ``KEYBOARD MENU'' which allow other forms of filtering.) The FILTER sub-menu is shown in Figure 25.

Figure 25: The FILTER Sub-Menu Commands

The available commands and their functions are:

EXIT: Leave FILTER sub-menu, and return to IMAGE PROCESSING (GENERAL) main menu.

EXCHANGE: transfers data from the ``main program array'' to the ``memory array'', and vice versa. This may be useful to save some data whilst another operation is overwriting the ``main program array''.

SMOOTH: Apply a top-hat smoothing function of specified size to the current region of interest. Top-hat smoothing is the convolution of the data with a rectangular region of uniform intensity (1.0 divided by the number of pixels in the rectangle). This is a low-pass filter, and will filter out noise and high spatial frequencies. You are prompted for X BLUR SIZE and Y BLUR SIZE. This defines the size of the top-hat function in the X and the Y directions. The original data is transferred to the memory, whilst the current data array contains the smoothed data after the operations is completed.

?: Help text with short explanation of commands.

FULL: Sets the extent of the ``region of interest (ROI)'' or ``active data region (ADR)'' to be the whole of the currently defined data. Note: the filtering commands only affect the current region of interest.

ZOOM-IN: allows a sub-region of the data to be selected and displayed. It should be noted that all operations only work on the current selected region.

HELP: Detailed help text on the interface and the available commands.

MEDIAN: Applies ``median'' filter of user specified size. Median filtering is replacing each data value by the median value in a rectangular region centred on the pixel. This is a non-linear filtering which filters out noise whilst tending to preserve edges and other sharp features better than top-hat smoothing. The user is prompted for the MEDIAN FILTER X-SIZE and the MEDIAN FILTER Y-SIZE. This defines the number of pixels used to define the rectangle around each pixel for the calculation of the median value. The original data is transferred to the memory, whilst the current data array contains the smoothed data after the operations is completed.