High Resolution DEM

[deepgis]

Is it possible to create high resolution DEM using CARTOSAT-1 Stereo pair in Agisoft software? If possible I need help from you all.

[Lurker]

why you want to process remote sensing image with UAV software?

UAV software lack function for remote sensing data, I dont think you can make precision DEM from it

 

better process it with ERDAS or ENVI software.

and yes with stereo pair image you can make DEM from it

 

here :

https://www.researchgate.net/post/How_do_you_generate_DEM_from_Cartosat-1_STEREO_data

[aminevsaziz]

thanks for the new info

[deepgis]

Respected Lurker,

You provided the solution, I was looking for.

Thanks to you.

[pasfans01]

yes it is possible, as long as 

 

• There is a proper overlap (approximately 80% frontal overlap and 60% side overlap).
• It is possible to describe the camera in terms of position (latitude/longitude/altitude) and parameters (sensor size, pixel resolution, focal length). The position mentioned in this prequisities cant be interpreted as georeference information, but the sensor position, which is usually stored in the imagery metadata

[Lurker]

so the telemetry data available in metadata? all the telemetry data for UAV proc?

[Art409]

Deepgis, I don't think tou could perform that in Agisoft.
(sorry pasfans01, but I'm disagree with you)

 

Cartosat-1 is a pushbroom system, not a framed one.
That means the camera scan the ground track line by line continuosly,
as spatial vehicle move foward in orbit.

 

So, in order to achive results you need camara position for each scan line,
more than 10.000 per scene, that means more than 20.000 cameras for agisoft
(you need 2 scenes)

 

Besides agisoft search for tie points usin images patchs,
and you only could get part of a single line as a patch.

 

Even if (with a miracle) you could make the software go,
I don't think you could get any acceptable results,
even with very low resolution

 

Use PCI, ENVI or Erdas. lot less painfull, lot better results
and the most important: proved solution

 

Art

[pasfans01]

@Art409 yes, I also know about the pushbroom/frame geometry differences, But What I am talking about is just workaround or trial and error wise

pushbroom scanners sure need certain tools which might be different from the frame camera dedicated tools, but why not trying it, 

I can get a DEM from Landsat-8 between Row side overlap which is reaching 10 percent of total swath width of landsat-8, but yes, it does needs additional processing which tedious and irritating 

[Art409]

Happy new year!. 

Agisoft solution is for frames.  1 frame, exterior orientarion (aka Camera in agisoft)

A landsat 7 scene got a aprox 180km along track, that menas the camera moved more than 180Km. ¿You took that as a simple scene?

 

 

 

 

[pasfans01]

personally I dont want to get limited by those theoretical stuffs, after all remote sensing is all about discovery, 

[Art409]

LOL.  Here goes one example of "theoretical stuffs". Why don't try to build a satellite instrument to see earth surface freflectance in ultraviolet o shoter wavelegnghs?

Answer: it''s IMPOSSIBLE to see that from space due atmospheric absortion!!!

What you call "theoretical stuffs" is the cause that we need the soft. "Discover" dosen't mean "ignore" knowledge.

Belive me, maybe the soft allow you to get some result. But these result is a useless.

I hope your doctor don't sahre your point o view!

 

 

 

[Lurker]

i like this discussion, keep it up 

we need more like this, 

BTW, I think the best approach is keep the theory as the corridor to discover a new things

try something new is good but there is a limitation based on theory, so let discuss how far that limit 

carry on 

[Art409]

Thanks Lurker!.

I don'y want to sound unpoilte. Sorry if my english limited me in that way pasfan01.

 

The key is that the soft, it's an implementation of some tehoretical algorithm.

Sometimes, we coud use or adapt that algorithm for a different prupose, but sometimes... is usefull.

Knowing the photogrammetric tehory, I advise that using Agisodt for pushbroom (or barridors like Landsat TM and ETM) CAN'T perform fine.

Usually in remote sensig DTM there is not not an inmediate risk for the user, But this sound to me like some one using an enginieering software designed for bridges structural analysys, trying to calculate the dome of a stadium, 

What I mean with "Can' perform fine"?? I state that the deliverable that you can obtain with cartosat-1 and Agisoft, will be less accurate that any ASTER DTM, OR SRTM (Ok SRTM data was taled in 2000, but Aster is still operative)

I'd out of spatial imgining for a couple of years, so I don't know what images are available today.

 

Art

 

PS: If any one want an explanation about why the movement of the camera (180km) is relevant, I can explain that

Edited January 8, 2017 by Art409
typing correction and PS