Hello GRASS users,
I am wondering if there is anyone out there using GRASS with non-geographical, spatial data? Specifically, the integration and rectification of raster and vector data which is NOT geographically referenced.
We are trying to build a 3D visualization application of gene expression data from the mouse cerebellum, from embryo through day 12 after birth. The gene expression data are spots from cross sectioned tissues, or "raster data" with no geographical reference/projection. These raster sections must be aligned (registered) with each other and then transformed (scaled, rotated, translated -- in GIS speak: rectified?) to fit into a 3D model (vector) containing surface meshes marking anatomical features of the cerebellum (exterior surface, cell layers, etc...).
Although I don't know much about the details of GIS or GRASS, it seems to me that GIS (and GRASS by extension) has handled this mix and match and rectification of raster and vector representations very well, but this functionality is based on defined and translatable geological projections and coordinate systems. Will the functionality work when the reference coordinate system is generic or home spun?
Are we swimming up stream if we try and stuff this data into GRASS or could it work? Your thoughts would be much appreciated.
(ASIDE: Isn't this always what happens in software? You build someone a shoe, a good shoe, for walking on land. Then, someone comes along and asks "Hey, I could walk on water with that shoe, couldn't I?" 
Thanks,
Carl Trapani
Skytop Software
Carl Trapani wrote:
Hello GRASS users,
I am wondering if there is anyone out there using GRASS with non-geographical, spatial data? Specifically, the integration and rectification of raster and vector data which is NOT geographically referenced.
We are trying to build a 3D visualization application of gene expression data from the mouse cerebellum, from embryo through day 12 after birth. The gene expression data are spots from cross sectioned tissues, or "raster data" with no geographical reference/projection. These raster sections must be aligned (registered) with each other and then transformed (scaled, rotated, translated -- in GIS speak: rectified?) to fit into a 3D model (vector) containing surface meshes marking anatomical features of the cerebellum (exterior surface, cell layers, etc...).
Although I don't know much about the details of GIS or GRASS, it seems to me that GIS (and GRASS by extension) has handled this mix and match and rectification of raster and vector representations very well, but this functionality is based on defined and translatable geological projections and coordinate systems. Will the functionality work when the reference coordinate system is generic or home spun?
Are we swimming up stream if we try and stuff this data into GRASS or could it work? Your thoughts would be much appreciated.
Several people have used GRASS with local archeology data, and others have used traditional (i.e. only locally referenced) cadastral data in GRASS, so I don't see why you shouldn't be able to use it.
AFAIK, you will need some sort of projection system to be able to use tools like i.rectify, but since you are working at an extremely large scale (sub-centimetric I would suppose), I actually suspect that you could probably use any arbitrary existing projection system as there would be no difference between them at that scale.
Moritz