The ROBERTS function returns an approximation to the Roberts edge enhancement operator for images:

where (j, k) are the coordinates of each pixel F_jk in the Image. This is equivalent to a convolution using the masks,

where the underline indicates the current pixel F_jk. The last column and row are set to zero.

Examples

---

If the variable myimage contains a two-dimensional image array, a Roberts sharpened version of myimage can be displayed with the command:

TVSCL, ROBERTS(myimage)

Syntax

---

Result = ROBERTS(Image)

Return Value

---

ROBERTS returns a two-dimensional array of the same size as Image. If Image is of type byte or integer, then the result is of integer type, otherwise the result is of the same type as Image.

Note: To avoid overflow for integer types, the computation is done using the next larger signed type and the result is transformed back to the correct type. Values larger than the maximum for that integer type are truncated. For example, for integers the function is computed using type long, and on output, values larger than 32767 are set equal to 32767.

Arguments

---

Image

The two-dimensional array containing the image to which edge enhancement is applied.

Keywords

---

None.

Version History

---

4.0	Introduced

See Also

---

EDGE_DOG, EMBOSS, LAPLACIAN, PREWITT, SHIFT_DIFF, SOBEL