Given an image consisting of black lines (a few pixels wide) on white background, what is a good way to find the coordinates along the the lines, say for every 10th pixel or so? I am considering using PIL for the task, but other python or java-based libraries would also be OK.
Ideally the coordinates would point to the middle of the line, but as the lines are narrow, it's enough that they point somewhere inside the line.
A very short line or a point should be identified with at least one coordinate.
Usually, Hough transformation is used to find lines. It gives you the parameters describing the line (which can be transformed easily between different representations), and you can sample this function to get your sample points. See http://en.wikipedia.org/wiki/Hough_transform and https://stackoverflow.com/questions/tagged/hough-transform+python
I only found this http://coding-experiments.blogspot.co.at/2011/05/ellipse-detection-in-image-by-using.html implementation in python, which actually searches for ellipses.
Related
I have several lines in a binary image. I know the code bridgeBW = bwmorph(closeBW, 'bridge'); will connect the lines if they are close enough, but so far I've only seen it do that in a one pixel range. Is there a way to increase the distance and bridge lines that are farther away?
I ended up using a line based strel method instead of one defined by a shape.
I would like to expand a line to a wider polygon. Add 10 meter on both sides of the line for example.
Here is an example of what I would like
Take this line
And expand it to a wider polygon, like this
I did this manually, is there a way to do this automaticly?
Changing the KML or using a program?
Thanks
Vincent
Depending on how accurate you need this – this is not trivial.
One possible algorithm could be:
for each segment do
expand segment to rectangle with width 2r
targetShape.join(rectangle)
next
for each point do
expand point to circle with radius r
targetShape.join(circle)
next
targetShape.outerHull(precision)
Each single line in this routine is tricky and depends on your expectations.
You could leave out the circles and instead make the rectangles longer, but this wil not work on sharp turns.
All of this involves ugly calculations to figure orthogonal lines etc.
You could try it in an graphic tool, gimp or inkskape :-)
I have a binary image with two white vertical segments separated by a small gap. I would like to calculate the distance between the two segments. Or better the gap.
My first attempt: find the profile of the two segments (using bwboundary and bwtraceboundary) and then find the intersection between this profile with horizontal line scanning the whole image. The number of lines without intersection represents the distance between the two segments.
I would like to find this gap without detecting the profile. Is there a way?
Thank you.
You can use measuretool from the MATLAB File Exchange by Jan Neggers to retrieve geometrical information of images.
I am creating a Microsoft Word document using the OpenXml library. Most of what I need is already working correctly. However, I can't for the life of me find the following bit of information.
I'm displaying an image in an anchor, which causes text to wrap around the image. I used WrapSquare but this seems to affect the last line of the previous paragraph as shown in the image below. The image is anchored to the second paragraph but causes the last line of the first paragraph to also indent around the image.
Word Screenshot http://www.softcircuits.com/Client/Word.jpg
Experimenting within Word, I can make the text wrap how I want by changing the wrapping to WrapTight. However, this requires a WrapPolygon with several coordinates. And I can't find any way to determine the polygon coordinates so that they match the size of the image, which is in pixels.
The documentation doesn't even seem to indicate what units are used for these coordinates, let alone how to calculate them from pixels. I can only assume the calculation would involve a DPI value, but I have no idea how to determine what DPI will be used when the user eventually loads the document into Word.
I would also be satisfied if someone can explain why the issues described above is happening in the first place. I can shift the image down and the previous paragraph is no longer affected. But why is this necessary? (The Distance from text setting for both Left and Top is 0".)
The WrapPolygon element has two possible child elements of LineTo and StartPoint that each take a x and y coordinate. According to 2.1.1331 Part 1 Section 20.4.2.9, lineTo (Wrapping Polygon Line End Position) and 2.1.1334 Part 1 Section 20.4.2.14, start (Wrapping Polygon Start) found in the [MS-OI29500: Microsoft Office Implementation Information for ISO/IEC-29500 Standard Compliance]:
The standard states that the x and y attributes are represented in
EMUs. Office interprets the x and y attributes in a fixed coordinate
space of 21600x21600.
As far as converting pixels to EMUs (English Metric Units), take a look at this blog post for an example.
I finally resolved this. Despite what the standard says, the WrapPolygon coordinates are not EMUs (English Metric Units). The coordinates are relative to the fixed coordinate space (21600 x 21600, as mentioned in the quote provided by amurra).
More specifically, this means 0,0 is at the top, left corner of the image, and 21600,21600 is at the bottom, right corner of the image. This is the case no matter what the size of the image is. Coordinates greater than 21600 extend outside the image.
According to this article, "The 21600 value is a legacy artifact from the drawing layer of early versions of the Microsoft Office."
When I make a figure in Matlab, with a legend and a rectangle that touches the y axis (strange, I know) upon exporting the figure to eps (or pdf) I've noticed that the rectangle obtains the line-style of the last line drawn (rather than what the rectangle was drawn with)
This behaviour also occurs for rectangles drawn after the one that touches the axis...
This doesn't happen if the rectangle is drawn before the legend is created....
Needless to say, it took me half a day to create a minimal example:
clf
L=plot(X,sin(X),'--');
legend(L,'sin(x)')
rectangle('position',[0.001,.1,.7,.7])
rectangle('position',[0,.5,.6,.7])
rectangle('position',[0.001,.3,.5,.7])
%legend(L,'sin(x)')
On the screen the 3 rectangle have solid lines, as they should. but once they are exported, the result has the last two with dashed lines (like the sin(x)). If the legend command is done later (as in the commented out line), everything works as it should....
Is this a feature or a bug?
This is not a feature. I am submitting this to development.
You found a workaround that works with minimal code gymnastics. I would document it in your code so someone does not change it unknowingly and move on.
If you are open to other output formats, notice this is not an issue with formats that use an output filter of MATLAB.
http://www.mathworks.com/access/helpdesk/help/techdoc/ref/print.html
(Graphic Format Files section, right column in table)
-Doug, Advanced Support at MathWorks dealing with graphical issues.