imagej MER basics - Page 11

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


Posts: 3465

Reply: 201



PostPosted: June 12, 2012 3:25 PM 

I forgot to mention why you didn't find this image on my Flickr site. For the first year or so I was hosting all my images on my own web site - but the job of maintaining and managing the site grew weary - so I switched to Flickr. Moving the early pictures to Flickr was one of those tasks I intended to do when the rovers died.

Until then if you have any favorites just post a reference on this topic and I will redo with original data - and use a lot of years of additional experience.

dx


Posts: 1661

Reply: 202



PostPosted: June 13, 2012 1:21 AM 

h>>>

Thanks for your 201. Yes, I too had a web site [twice with the same IP] in those early days but the IP closed all addresses and never went back into it.

yt
dx

dx


Posts: 1661

Reply: 203



PostPosted: June 15, 2012 9:30 AM 

h...
LG 1080p died yesterday will be a while until I present more images. Checking out LED's today.
yt
dx

dx


Posts: 1661

Reply: 204



PostPosted: June 16, 2012 3:46 AM 

h>>>

OK, new LED LCD ASUS 50M:1 24" monitor...trying out some color works on old Spirit Sol 243. Not in your flickr h!

yt
dx

hortonheardawho


Posts: 3465

Reply: 205



PostPosted: June 16, 2012 11:57 AM 

Spirit sol 243 was pre-Flickr. The first Flickr picture was uploaded on Dec 4, 2004 ( Spirit sol 327 - Oppy sol 307 - my 60th birthday ).

I located the original image that I did of the sol 243 image - but didn't like it ( too "yellow" ) and decided to redo it using original data:

sol 243 ( Sep 8, 2004 ) L234567 saturated false color of unnamed rocks:

The twist on this image is the saturated false color.

The basic idea is to process the image in the HSB color space.

The first step is to change the type of an RGB color image to a HSB Stack by using the imagej command Image / Type / HSB Stack. The second slice in the stack is the image color saturation.

My "satsqt" processing steps are:

1) make a duplicate of the saturation slice,
2) convert to 16 bit,
3) multiply by 255,
4) apply the square root function,
5) convert to 8 bit,
7) copy back to the saturation slice,
Cool convert the HSB stack back to an RGB image using the command Image / Type / RGB Color.

The "a" processing was a final color and brightness "tweak" ( I thought the final was still too yellow ) .

OK, now what do we see. Although I think the dark rock is what attracted the rock guys, what interest me more is the "green" areas.

Of course, there real-l-ly isn't any green on Mars - but these areas are more green than most of the surrounding areas. That is in fact a - er - fact. So what is responsible for the color of these areas?

As you might guess, I have lots of questions and damn few answers about Mars.

dx


Posts: 1661

Reply: 206



PostPosted: June 16, 2012 4:40 PM 

h>>>

Excellent info in 205. Thank you for the 'HSB color space' link...fulfills my day...you can seemingly locate the most relevant sites I have never thought of to find for use in color imagery...many thanks, again.

I have never doubted your presentations or possession of imagery quality from the MER cameras.

All the very best h,

yt
dx
BTW>>> you have 2 years on me. I will never catch up. LOLOL

hortonheardawho


Posts: 3465

Reply: 207



PostPosted: June 16, 2012 5:19 PM 

dx, as a door manufactured by the Sirius Cybernetics Corporation would say I am "Glad to be of service".

You do realize that the "homework" is now a saturated MER color image. The soils around Spirit's final resting place are a good place to start.

dx


Posts: 1661

Reply: 208



PostPosted: June 16, 2012 7:36 PM 

h>>>

Literally speaking, in design I hate doors-always have, always will. My work NEVER accompanies doors. Wifey doesn't get it, but who is she anyway...I'm from somewhere else. Architecture is thought-only some of us have it!

yt
dx

dx


Posts: 2

Reply: 209



PostPosted: June 17, 2012 12:48 AM 

h>>>

Been thinking about your comment in 205. I also saw the 'green' on the right side of the image dispersed amongst the rocks, tried to remove it to get at its natural state as close as possible, but it would not remove itself. SOOO, the dark rock, in my mind, must have been a flaming meteorite that may have caused a fire of sorts at this area. Maybe the Martian dust is flammable at high temperatures! If the colors represent a reasonable image for these rocks at this place, then some thing happened here not too long ago, because all the rocks are missing the dusting cover that seems to consume everything on Mars. The 'green'>>>I don't know either, but the 'black' rock resting on a typical Mars rock does have a splash of 'red' on its side facing us. Martian growth, or impact mark or none of the above?

yt
dx

dx


Posts: 1661

Reply: 210



PostPosted: June 21, 2012 12:40 PM 

h>>>

I'll look at the final resting place as you mentioned in your 207, but I wanted to take a look at the first pics from Spirit when it landed. There are 3 of them, which I will post here. I will also try the '16-bit-sqrt' on this image to see if there are any differences.

yt
dx

hortonheardawho


Posts: 3465

Reply: 211



PostPosted: June 30, 2012 11:28 AM 

sol 2981 ( Jun 13, 2012 ) montage of "maxRGBsqt" image processing:

see Flickr comments for an explanation of the processing steps.

Mizar


Posts: 692

Reply: 212



PostPosted: June 30, 2012 1:04 PM 

How advanced this thread suppose to be then Wink BTW great work and very informative!

dx


Posts: 1661

Reply: 213



PostPosted: July 7, 2012 8:42 AM 

h>>>

I'd like to thank you here in this thread for your first post on April 15, 2012 to help people like me gain your insights into producing images using 'imageJ'. Over the past 3 months I found the process very exciting and will continue to discover the many aspects of watching your downloads here and on your flickr. Your caring for me and answering my many questions has brought me a new adventure in color processing the pictures from Mars. I may have more 'Q's' in the future you may be able to answer, but I have read all your tips and lead site links which have been invaluable to me.

I will continue to post imageJ composites I find interesting. They may not contain questionable organic rock structures as you have conditioned your mind's eye to see, but perhaps I too may see something of interest in the processed images to come.

All the very best in your personal health condition and many many thanks again. I am not leaving the MRB, just wanted to give you my praise at this time for your endless help.

All the very best 'h' and trust you have completely recuperated from last years' malady.

yt
dx

dx


Posts: 1661

Reply: 214



PostPosted: July 10, 2012 11:27 AM 

h>>>

This is help time. I have researched the net looking over the years for something that would convert a B/W photo into a colored picture. PSP and imageJ can split the channels but when merging them they retain the B/W image.

Is there some kind of procedure I am missing on the R-G-B splits before a merge?

Appreciate your help, thanks.

yt
dx

hortonheardawho


Posts: 3465

Reply: 215



PostPosted: July 10, 2012 12:51 PM 

er, dx, as far as I know, there are no totally automated procedures to colorize B&W images without human input - yet.

There are obviously computer tools for colorizing B&W ( there are lots of colorized classic B&W movies ) - but all of them require guesses about the color of regions in the image - usually input by a human.

Some of the more clever programs can do a scene analysis ( this area is a face - this area is a tree - this area is a dog - this area is a rose - etc... ) and from a feature - color table ( created by humans? ) can automatically assign color.

Strangely, the job of converting 2d images to 3d is easier than coloring B&W images - and there are programs that do a decent job of converting simple 2d images to 3d - especially films. Some are actually freely available on the web ( see G'MIC. )

I played around a few years ago with colorizing navcam images - based on analyzing pancam brightness of the same area - but decided it was not helpful in understanding the images.

What B&W images did you have in mind to color?

"Someday" ( maybe tomorrow at 3PM? ) computer programs will exist to fully extract all the feature information in a movie, create 3D color models of all the features ( including the actors and their voices ) and then allow a new script to be input to create an entirely new 3D color movie!

Maybe James Cameron is working on the idea already.

dx


Posts: 1661

Reply: 216



PostPosted: July 10, 2012 3:50 PM 

h>>>

Many thanks for your response. Checked out your supplied links. Don't have gimp but did find many years ago a program -free- called BlackMagic. You can not save your colored image work [but there are ways for that] and one does have to color by eye and mind control. Its not what I was looking for but retained the program for inspiration coloring some 'old' B/W photos from yesteryear. Some look very good colored and you have to understand colors from the time and what and how to match them up to make a good refined final product. There are various colors to choose from with given name selections-like 'lips, eyes and skin tones'. That is OK-but really not what I was looking for.

If I channel split a B/W into RGB and then color each RGB with their own color designation, then merge the colored results-usually get BLACK!!! I thought there would be a process that I missed like-mathematics or adding arithmetic to each channel. I have not tried these things out to any extent but was thinking you may have tried them yourself along the way to imageJ.

yt
dx

hortonheardawho


Posts: 3465

Reply: 217



PostPosted: July 18, 2012 3:27 PM 

sol 81 ( Apr 16, 2004 ) example of L257 "natural color" vs enhanced difference false color processing:

using RAD data.

The basic idea behind enhanced difference processing is the color map:

( R, G, B ) = ( G+4*(R-G), G , G+4*(B-G))

This works extremely well for channels with small differences between the filters ( such as the infrared filters )

Now, does anyone "see" why micro-channels are NOT just "smooth dust areas"?

LWS


Posts: 3062

Reply: 218



PostPosted: July 18, 2012 11:29 PM 

Hort; re. your #217 above. Could you do the same processing for the small crater in sol 88, perhaps also the soil digging in the same sol?

The results may surprise even you.

Winston.

hortonheardawho


Posts: 3465

Reply: 219



PostPosted: July 19, 2012 1:23 PM 

sol 88 ( Apr 23, 2004 ) L257 G+4xd false color of Fram crater:

and L257 G+4xd of rover tracks near Fram:

from RAD data.

LWS


Posts: 3062

Reply: 220



PostPosted: July 19, 2012 2:24 PM 

Hort; Thanks!

I should have more clearly identified the images that I thought might be special.

Re. the crater, the middle image in the crater series. The one next to the one you processed. It shows what looks like pale tracks along the crater edge moving over a very strange looking rock. resembles a large napoleonic hat.

re. the tracks, the earlier image with the prominent rind or impact meld, or evaporite fragments or whatever. Unfortunately this one does'nt have R filter images for 3D.

My apologies and grateful if you would do them.

Winston


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