reprojector distorting arcs

Hi, 

 

 

I have no experience with this particular case, but have you tried out the two possibilities of reprojecting?

1. the transformers to do so
2. defining the Mercator projection in the output of your writer?

 

During the FME World Tour, I was informed that the first one is to be preferred to the second method. It used to be the other way around before.

 

 

 

best regards, 

 

Jelle

 

 

If I reproject the file prior to processing through FME to stroke the arcs, the areas are not distorted.  If I use FME to reproject and then stroke the arcs you can see the distortion.  I haven't been able to successfully reproject the files without this distortion. 

 

 

I also noticed that if I run the original features through an ArcPropertiesExtrator prior to reprojection, the primary and secondary radius are equal.  Once the arcs are reprojected the primary and secondary radius are no longer equal. 

 

Again, if I start with the file that was reprojected outside of FME then the primary and secondary radius are equal.  

 

 

I'm attaching an image that shows what I am seeing.  The blue line is shows the area from the file that was reprojected prior to FME.  The brown line shows the area that was reprojected within FME.  As you can see the longer arc segments are heavily distorted.

 

 

 

That sounds peculiar. The reprojections I've done so far, have always worked out well.

Hi,

 

 

As you know, the same longitude and latitude values (in degree) are different in the actual distance (in meter, mile etc.). Every shape will be distorted naturally if you project them to Mercator from LL. It's a common result on not only arc but any other shape - e.g. line, rectangle, ellipse etc.. If the original arcs are representing correct shapes of features, I think the distorted shapes (stroked before projection) are rather correct than not distorted shapes. -- in my personal opinion.

 

 

Takashi

"As you know, the same longitude and latitude values (in degree) are different in the actual distance (in meter, mile etc.). Every shape will be distorted naturally if you project them to Mercator from LL. It's a common result on not only arc but any other shape - e.g. line, rectangle, ellipse etc.. If the original arcs are representing correct shapes of features, I think the distorted shapes (stroked before projection) are rather correct than not distorted shapes. -- in my personal opinion."

 

 

I could see this except for the fact that the arcs reprojected in FME are different than the arcs when they are reprojected in other software.  As shown in the image above.  When the same file is reprojected outside of FME and then stroked I get the correct result (shown in blue).  When the file is reprojected within FME and stroked I get the incorrect result (shown in brown).  And again, the distortion is greater the longer the arc segment.  As you can see in the image only the two longer ones are distorted.  The shorter segments match.  

I should also add, that during this reprojection the arc segments are not moving correctly when compared to other features.  For example, on the topmost arc segment, prior to reprojection and stroking it was completly on the southern side of a linear boundary.  After the reprojection and stroking, it now crosses that linear boundary.  If it was correctly reprojecting it's spatial relationship to other features should not change. 

The brown arcs in the image are indeed strange. But I guess the reason is that the arcs were not stroked before projection.

 

As I mentioned before, I think "the distorted shapes (stroked before projection)" will be the closest shapes to the actual features. How is the result, if you stroked them before projection?