imagej MER basics - Page 13

Previous 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 Next
Author Message
hortonheardawho


Posts: 3465

Reply: 241



PostPosted: August 29, 2012 5:32 PM 

jj, thanks for your suggestions. It inspired me to write a new version of the demosaic code.

The basic idea is to imagine a "central" grid which is between the Bayer grid and then use the nearest R,G1,G2,B filter values for the central grid. The programming will amuse you.

There still seems to be a problem with color banding - but I can't decide if the banding is a "feature" or a "bug".

It looks like the practice of crippling and degrading the data will continue, so I will continue hammering away at the problem. I still have a few ideas left...

hortonheardawho


Posts: 3465

Reply: 242



PostPosted: August 29, 2012 11:34 PM 

It turns out the banding was a "feature". Nathan Yan, the owner of the Demosaic site informs me that the lighthouse image was not an original 100% size - and that the color banding ( Moire patterns ) can result from resizing.

The above test - also from his site - works out much better.

I wonder if NASA is doing something nasty - like resizing? Naaah. JPGing Bayer patterns is nasty enough.

b0nafide


Posts: xxx

Reply: 243



PostPosted: August 31, 2012 5:06 PM 

Hey, maybe join forces with the charismatic and intelligent blogger here to get JPL to do things differently. [link]

hortonheardawho


Posts: 3465

Reply: 244



PostPosted: August 31, 2012 5:49 PM 

Hi son.

I think Emily is co-opted. She has access to people and information that she would most likely lose if she raised any embarrassing questions for NASA.

I think I will spend my remaining days taking pretty pictures.

I have been reviewing various Nikon digital cameras in the last week and will select one within the next few days. Then I will be off to the races learning to make beautiful pictures once again.

I like the idea of the availability of the raw 12 bit Bayer files as "digital negatives" for editing - so the work I have been doing with Bayer demosaic code will be put to good use.

How's your WiFi helicopter project coming along?

dx


Posts: 1661

Reply: 245



PostPosted: August 31, 2012 6:31 PM 

h>>>

That's the second time you have mentioned 'remaining days'...WTF is going on?

...respectfully,
yt
dx

b0nafide


Posts: xxx

Reply: 246



PostPosted: August 31, 2012 6:45 PM 

Oh the helicopter is on the backburner, waiting for parts.

Like your other son has said to me: "Never attribute to malice what can be explained by simple incompetence" or something like that.

Maybe in all the excitement, some network admin is running some batch processing and doesn't even realize what they are doing to the scientific payload?

b0nafide


Posts: xxx

Reply: 247



PostPosted: August 31, 2012 6:47 PM 

ie. maybe the public website isn't run by scientists and JPL would be okay with the uncompressed data being public?

hortonheardawho


Posts: 3465

Reply: 248



PostPosted: August 31, 2012 7:37 PM 

Perhaps incompetence on just one group's operational level would explain the problem - but there is an across-the-board unreasonable degrading of all the camera data - and by all the organizations involved - from JPL to Malin to LANL to NASA. Even the NASA Photojournal site releases only 8 bit TIFs as "Full Resolution" images.

The only source of "real" Mars Rover data is the Analyst's Notebook.

As for accessing the "real" data in "real" time...

At the start of the MER mission I figured out exactly which server and which directories contained the real, live data.

I also discovered that the directories were not secure, so I downloaded ONE Navcam pair of pictures, processed a 3D and posted it - and almost immediately received an e-mail from a dude I didn't know wanting to know how I made the 3D. I explained exactly how to access the live data - and the very next day the server password challenged me.

So yes, "they" care very much about protecting the "live" data.

dx, don't get too excited. I'm not going anywhere real soon. I'm just changing my priorities from bashing my head against a wall that refuses to move to making pretty pictures.

dx


Posts: 1661

Reply: 249



PostPosted: August 31, 2012 8:12 PM 

h>>>

Thanks, 'h'...all the best with your camera hunting challenge.

yt
dx

Barsoomer


Posts: 344

Reply: 250



PostPosted: August 31, 2012 8:14 PM 

Horton, please at least use the 6-month old Analyst's Notebook data. I notice they have APXS data on Amboy (at Greeley Haven) now, but I unfortunately I don't know how to interpret the Alpha and X-ray graphs.

b0nafide


Posts: xxx

Reply: 251



PostPosted: August 31, 2012 10:56 PM 

Well as far as protecting the live data, I think a sensible network policy that prevents tampering with the data is certainly reasonable and I can understand a password requirement for write access anyways.

Maybe they don't know how to allow anonymous read access without allowing anonymous write access and it has therefore become somebody's job to move the data to the public face of NASA. I can vividly imagine a web design company taking on the task and not giving a second thought to compressing graphics before posting them, if there isn't already an 'optimize everything' feature in their IDE...

b0nafide


Posts: xxx

Reply: 252



PostPosted: August 31, 2012 11:04 PM 

Anyways if read access is all that you had before, perhaps somebody is panicking because they don't understand that there are no security implications to just having read access.

Are the recent high res images for sale?

MPJ


Posts: 250

Reply: 253



PostPosted: September 1, 2012 1:21 PM 

Barsoomer, have you got a link to the APXS data of Greeley Haven targets at hand? I really would like to have a look too.

Barsoomer


Posts: 344

Reply: 254



PostPosted: September 1, 2012 3:06 PM 

link

More info here:

link

Sorry if this is OT for this thread. Horton, please move this if it belongs elsewhere.

dx


Posts: 1661

Reply: 255



PostPosted: September 3, 2012 5:18 PM 

h>>>

With all due respect, have you updated your 'Bayer-RGB'? I have noticed in your 241, left pan landscape image, that the green squares are gone, yet are retained on the right image. I have uploaded most of Curiosity's landscape images and used your Bayer-RGB only to have green accents present in the hills.

Thanks.

yt
dx

hortonheardawho


Posts: 3465

Reply: 256



PostPosted: September 3, 2012 5:49 PM 

dx, the latest version is in the comments of this Flickr image:

The JPG artifacts are still there - especially in low contrast images like views of Mount Sharp.

I am waiting to see what the MSL site does with color images. If they continue to post the degraded Bayer images then I will try a couple of tricks I think might mitigate the problem. But be forewarned: there ain't no free lunch. The cost will be reduced resolution.

They may gave already done some tweaking of their code. This "raw" Bayer image from under the rover came out quite nicely.

hortonheardawho


Posts: 3465

Reply: 257



PostPosted: September 3, 2012 6:12 PM 

The latest right Mastcam-100 image was posted as a "regular" JPG with decent color and resolution - so perhaps NASA has heard the screams across the land and will only post "real" pictures.

BTW, there is *still* a slight green bias in the image. I had to "stomp" on a patch of green ( color balance to grey ) in the upper right corner.

dx


Posts: 1661

Reply: 258



PostPosted: September 3, 2012 6:41 PM 

h>>>

I just tried out Bayer-new versus Bayer-RGB on several images...WOW! Much better performance for which I thank you. I did add 1 'sharpen' tweak to the final image. Can you add that tweak as an extra step in your program, or is it better to add it at the end? [besides that I wouldn't fool around with your program for I wouldn't know where to add that sharpen note nor the wording!]

As an added note and I am sure and other agree,...those damned little thumbnails drive me crazy...OK, OK I'll calm down, but for 2.3B$ just to have thumbs...I don't believe its come to that! Get a human to Mars with a brand new Canon! LOLOL

yt
dx

hortonheardawho


Posts: 3465

Reply: 259



PostPosted: September 5, 2012 2:28 PM 

For those who might be interested, I have ordered a Finepix HS30EXR digital camera - which should arrive in the next few days.

It has a zillion features, and I will be exploring all of them in preparation for a trip in October.

How could I not buy a camera that offers in camera panoramas, HDR, 3D, raw format output, super macro, 1920x1080 HD 30 fps plus slow motion movie modes - all seen through a 16 Megapixel EXR CMOS sensor and a 30X optical zoom lens coupled with advanced image stabilization for those hand held telephoto and low light shots!

I don't believe the claimed ISO speed of 12,800. Maybe I will be able to photograph that black cat at midnight.

It has a zillion program modes - plus full manual control - including manual white balance with a gray card.

There are some quite bizarre features that I will have to see to believe.

Like 8 individual face recognitions that you can program - and dog or cat face recognitions! I'm not real sure what possible use that will have - but I will give it a try.

The HS30EXR camera manual is 146 pages.

I plan on posting test pictures as I create them - so be prepared for a lot of weird pictures in my Flickr photostream in the coming weeks.

dx


Posts: 1661

Reply: 260



PostPosted: September 6, 2012 7:51 AM 

h>>>
Congrats on your selection. Very nice camera. Read all about it. Take some good ones!
yt
dx

Previous 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 Next


Join the conversation:















Very Happy Smile Sad Surprised
Shocked Confused Cool Laughing
Mad Razz Embarassed Crying or Very Sad
Evil or Very Mad Twisted Evil Rolling Eyes Wink
Powered by MTSmileys