NAME
Image::Leptonica::Func::edge
VERSION
version 0.04
edge.c
edge.c
Sobel edge detecting filter
PIX *pixSobelEdgeFilter()
Two-sided edge gradient filter
PIX *pixTwoSidedEdgeFilter()
Measurement of edge smoothness
l_int32 pixMeasureEdgeSmoothness()
NUMA *pixGetEdgeProfile()
l_int32 pixGetLastOffPixelInRun()
l_int32 pixGetLastOnPixelInRun()
The Sobel edge detector uses these two simple gradient filters.
1 2 1 1 0 -1
0 0 0 2 0 -2
-1 -2 -1 1 0 -1
(horizontal) (vertical)
To use both the vertical and horizontal filters, set the orientation
flag to L_ALL_EDGES; this sums the abs. value of their outputs,
clipped to 255.
See comments below for displaying the resulting image with
the edges dark, both for 8 bpp and 1 bpp.FUNCTIONS
pixGetEdgeProfile
NUMA * pixGetEdgeProfile ( PIX *pixs, l_int32 side, const char *debugfile )
pixGetEdgeProfile()
Input: pixs (1 bpp)
side (L_FROM_LEFT, L_FROM_RIGHT, L_FROM_TOP, L_FROM_BOT)
debugfile (<optional> displays constructed edge; use NULL
for no output)
Return: na (of fg edge pixel locations), or null on errorpixGetLastOffPixelInRun
l_int32 pixGetLastOffPixelInRun ( PIX *pixs, l_int32 x, l_int32 y, l_int32 direction, l_int32 *ploc )
pixGetLastOffPixelInRun()
Input: pixs (1 bpp)
x, y (starting location)
direction (L_FROM_LEFT, L_FROM_RIGHT, L_FROM_TOP, L_FROM_BOT)
&loc (<return> location in scan direction coordinate
of last OFF pixel found)
Return: na (of fg edge pixel locations), or null on error
Notes:
(1) Search starts from the pixel at (x, y), which is OFF.
(2) It returns the location in the scan direction of the last
pixel in the current run that is OFF.
(3) The interface for these pixel run functions is cleaner when
you ask for the last pixel in the current run, rather than the
first pixel of opposite polarity that is found, because the
current run may go to the edge of the image, in which case
no pixel of opposite polarity is found.pixGetLastOnPixelInRun
l_int32 pixGetLastOnPixelInRun ( PIX *pixs, l_int32 x, l_int32 y, l_int32 direction, l_int32 *ploc )
pixGetLastOnPixelInRun()
Input: pixs (1 bpp)
x, y (starting location)
direction (L_FROM_LEFT, L_FROM_RIGHT, L_FROM_TOP, L_FROM_BOT)
&loc (<return> location in scan direction coordinate
of first ON pixel found)
Return: na (of fg edge pixel locations), or null on error
Notes:
(1) Search starts from the pixel at (x, y), which is ON.
(2) It returns the location in the scan direction of the last
pixel in the current run that is ON.pixMeasureEdgeSmoothness
l_int32 pixMeasureEdgeSmoothness ( PIX *pixs, l_int32 side, l_int32 minjump, l_int32 minreversal, l_float32 *pjpl, l_float32 *pjspl, l_float32 *prpl, const char *debugfile )
pixMeasureEdgeSmoothness()
Input: pixs (1 bpp)
side (L_FROM_LEFT, L_FROM_RIGHT, L_FROM_TOP, L_FROM_BOT)
minjump (minimum jump to be counted; >= 1)
minreversal (minimum reversal size for new peak or valley)
&jpl (<optional return> jumps/length: number of jumps,
normalized to length of component side)
&jspl (<optional return> jumpsum/length: sum of all
sufficiently large jumps, normalized to length
of component side)
&rpl (<optional return> reversals/length: number of
peak-to-valley or valley-to-peak reversals,
normalized to length of component side)
debugfile (<optional> displays constructed edge; use NULL
for no output)
Return: 0 if OK, 1 on error
Notes:
(1) This computes three measures of smoothness of the edge of a
connected component:
* jumps/length: (jpl) the number of jumps of size >= @minjump,
normalized to the length of the side
* jump sum/length: (jspl) the sum of all jump lengths of
size >= @minjump, normalized to the length of the side
* reversals/length: (rpl) the number of peak <--> valley
reversals, using @minreverse as a minimum deviation of
the peak or valley from its preceeding extremum,
normalized to the length of the side
(2) The input pix should be a single connected component, but
this is not required.pixSobelEdgeFilter
PIX * pixSobelEdgeFilter ( PIX *pixs, l_int32 orientflag )
pixSobelEdgeFilter()
Input: pixs (8 bpp; no colormap)
orientflag (L_HORIZONTAL_EDGES, L_VERTICAL_EDGES, L_ALL_EDGES)
Return: pixd (8 bpp, edges are brighter), or null on error
Notes:
(1) Invert pixd to see larger gradients as darker (grayscale).
(2) To generate a binary image of the edges, threshold
the result using pixThresholdToBinary(). If the high
edge values are to be fg (1), invert after running
pixThresholdToBinary().
(3) Label the pixels as follows:
1 4 7
2 5 8
3 6 9
Read the data incrementally across the image and unroll
the loop.
(4) This runs at about 45 Mpix/sec on a 3 GHz processor.pixTwoSidedEdgeFilter
PIX * pixTwoSidedEdgeFilter ( PIX *pixs, l_int32 orientflag )
pixTwoSidedEdgeFilter()
Input: pixs (8 bpp; no colormap)
orientflag (L_HORIZONTAL_EDGES, L_VERTICAL_EDGES)
Return: pixd (8 bpp, edges are brighter), or null on error
Notes:
(1) For detecting vertical edges, this considers the
difference of the central pixel from those on the left
and right. For situations where the gradient is the same
sign on both sides, this computes and stores the minimum
(absolute value of the) difference. The reason for
checking the sign is that we are looking for pixels within
a transition. By contrast, for single pixel noise, the pixel
value is either larger than or smaller than its neighbors,
so the gradient would change direction on each side. Horizontal
edges are handled similarly, looking for vertical gradients.
(2) To generate a binary image of the edges, threshold
the result using pixThresholdToBinary(). If the high
edge values are to be fg (1), invert after running
pixThresholdToBinary().
(3) This runs at about 60 Mpix/sec on a 3 GHz processor.
It is about 30% faster than Sobel, and the results are
similar.AUTHOR
Zakariyya Mughal <zmughal@cpan.org>
COPYRIGHT AND LICENSE
This software is copyright (c) 2014 by Zakariyya Mughal.
This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.