Opened 16 years ago

Closed 16 years ago

#2510 closed defect (invalid)

Artifacts appear during reading raster with different scales

Reported by: omidsakhi Owned by: warmerdam
Priority: normal Milestone:
Component: GDAL_Raster Version: unspecified
Severity: normal Keywords:
Cc:

Description

sometimes during reading a raster band (I have worked with PCIDSK and ERDAS IMAGINE) when the rasterio should take care of the scaling, some horizontal lines appears as artifacts and the data in these lines have zero value in my case. by changing the scale, the location of the lines changes. only the one to one mapping is always without these kind of errors.

an image is attached. the attaches image is a shaded relief image from a large erdas imagine file format. the red lines are not suppose to exist and in fact by changing the scale (zooming in or out) they might disappear or the location might change.

Attachments (2)

erdas_imagine_art.jpg (37.2 KB ) - added by omidsakhi 16 years ago.
an example of artifacts
kalabast_2000.rar (475.3 KB ) - added by omidsakhi 16 years ago.

Download all attachments as: .zip

Change History (6)

by omidsakhi, 16 years ago

Attachment: erdas_imagine_art.jpg added

an example of artifacts

comment:1 by Even Rouault, 16 years ago

Could you provide a dataset or link to a dataset, and a code snippet that trigger the problem you're raising ?

by omidsakhi, 16 years ago

Attachment: kalabast_2000.rar added

comment:2 by omidsakhi, 16 years ago

I have attached the daraset. but the codes inare part of a software that needs lots of libraries including QT, OpenGL, CG , CGAL and ... and therefore the program doesn't start do easily. just open the file using

poDataset->GetRasterBand(1)->RasterIO(GF_Read,0,0,8497,6121,poData,725,522,GDT_Float32,0,0);

and you will see two horizental lines as artifacts at the top of the data. the first line start from about (349,5799) to (1356,5799) Cartesian coordinate and the second one start from (4868,6022) to (5875,6022).

please confirm that I am not paranoia.

Sincerely

comment:3 by omidsakhi, 16 years ago

sorry ! the coordinates of the lines are calculated relative to the original raster x and y size

comment:4 by Even Rouault, 16 years ago

Resolution: invalid
Status: newclosed

Well, it's not a bug, but a feature. Your source Imagine dataset has many small white strips. I've verified it with other viewers such as ERViewer. The fact that the strip position changes according to the scale is due to neirest neighbour approximation I think.

So I'm closing as invalid.

Note: See TracTickets for help on using tickets.