Hey Jason,

Geeky was good.

The method you described below seemed to work well. I unrotated the 
text and created the bounding box for each text element. However, 
for some reason the length between the min x and max x of the 
bounding box seemed only to be half the length of the text element. 
So I had to multiply the value of  (max x - min x) for each by 2  
and then add it to the original min x to give a new max x which made 
the bounding box roughly accurate.

I then intersected the bounding boxes with each other. Those with 
more than 1 _segment overlap were passed out via a tester. Thus all 
those bounding boxes with no overlaps were sent to a featuremerger 
which matched up the text element (straight from the orginal DGN) 
with the bounding boxes using the idgs_graphic_group attribute that 
was in the original DGN (the bounding boxes were the requestor, the 
text elements the supplier and text elements that were taken into 
the new dgn were taken from the referenced port).

That seemed to work. It did remove anything with an overlap...which 
is not exactly ideal but I'm still coming to grips with looping 
transformers so I might be able to refine the process just to remove 
one of the intersection text elements.

Thanks for your help once again.

--- In [email protected], "Jason Birch" <[EMAIL PROTECTED]> wrote:
>
> Sure :)
>  
> I don't have FME where I'm at (I'm in Vancouver for some C# 
training - yay), so sorry for vagueness on transformer names.
>  
> It would be nice if you could just force the geometry to a 
polygon, but I don't think that's possible.  However, it's a little-
known fact that you can use the bounding box transformers to get 
estimated bounding boxes for text elements.
>  
> If you have rotation in your text elements, you may need to 
unrotate them around their centroid, generate their bounding box as 
a polygon, and then re-rotate that around its centroid. 
>  
> Once you have the polygon in place, you can perform a spatial 
intersection.  You may want to do this in a loop within custom 
transformer, overlaying the polygons, removing the ones with the 
highest number of intersections with other polygons, passing the 
ones with zero intersections directly to the output, and running the 
remainder through a hidden input at the beginning of the custom 
transformer.
>  
> Mmm.  Maybe there are less geeky ways of doing this.
>  
> Jason
> 








Get the maximum benefit from your FME, FME Objects, or SpatialDirect via our 
Professional Services team.  Visit www.safe.com/services for details. 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/fme/

<*> To unsubscribe from this group, send an email to:
    [EMAIL PROTECTED]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 



Reply via email to