a problem about transformation for WGS84 to NAD83

[xicheng]

I got some tm images which has GCS_WGS_1984??and i want to reproject to NAD_1983 to fit the product of NLCD.i pick up the "project raster" tool and using NAD_1983_To_WGS_1984_5 which is suitable for North America area. But there had some pixels distortion which i attached at end. 
Could someone help me about this problems and how to imply the correct transformation?

Thanks,
Xi

original wgs84

nad83 after transfer

[JimW1]

My guess is that your raster is being rectified when it is being projected. I am assuming your WGS84 image was georeferenced but not georectified - this is typical for orthos from vendors. So when projecting to NAD83, Arc must georectify the image (Resampling Technique). The pixels are now all oriented north.

If this is the case, I unfortunately do not have any good solutions for you.

One not great idea would be to make the Output Cell Size to about half of what it is now and that would reduce the distortion but quadruple the storage space.

In the past I have had to request NAD83 world files for an ortho delivery from the vendor to avoid this rectification issue (and to save HDD space).

[xicheng]

Jim

thank you very much and i'm sorry for replying so late!

my images are landsat tm5 which are free provided at USGS's website,i am not sure whether the they were georectified.

i will try your method to half the cells and find whether it have a acceptable result.

thank you~

[MelitaKennedy]

I'm sorry, I should have posted about this earlier. Make sure the tool (or geoprocessing settings) is using nearest neighbor.

And, more importantly, the cell size is 30m. The difference between WGS84 and NAD83 in the contiguous US (48 states) is about a meter. I wouldn't bother reprojecting the raster. Use its property page in ArcCatalog or the Define Projection tool to set it to NAD83.

Melita

[xicheng]

thank you very much, Melita

actually, i found the problem is caused by that i choose the nearest neighbor reampling method  instead of binary method which can product nice result.

i will use your way in the future, thanks

Xi

[MelitaKennedy]

Oops, I had the resampling methods backwards! Thanks for correcting me, and I'm glad that you solved your problem.

Melita