PJ15_19,20,21,22,23


From Brian Swift's pipeline

PJ15_22 processed with updated Mathematica pipeline available at GitHub which includes flat-field/debias correction and new color balance.

Nice job Brian, I'm clearing out some hard drive space for the script update.

Here is my take on this awesome output...


Aligned rgb channels & image process. ( & flipped it )

PJ15 drafts, part 1.
Reprojected RGBs, part 1.

My upload bandwidth to missionjuno was rather narrow most of the night. So, some delay.

Here the reprojected RGBs in JPG:
#23:
Click to view attachment

#22:
Click to view attachment

#21:
Click to view attachment

#20:
Click to view attachment

#19:
Click to view attachment

#18, and #17:
Click to view attachment Click to view attachment

#15, #11
#10, #09
#08, #07,
and #06:
Click to view attachment Click to view attachment
Click to view attachment Click to view attachment
Click to view attachment Click to view attachment
Click to view attachment

Brian, the large image you put inline makes this page load very slowly.

Btw., this is the -- technical -- description I've added to the missionjuno upload:

The image is reprojected according to a slightly modified, and still preliminary geometrical camera model, cleaned from most repetitive camera artifacts and from most energetic particle hits, approximatly illumination adjusted, and stretched with gamma=4 with respect to radiometric values.
Applied camera geometry: pinhole focal length 1482.5 (square) pixels, Brownian radial distortion parameters K1=-6.0E-8, K2=3.0E-14, CX=809, CY=610.01, for all filters, relative to system with the following lower edge of readout regions: red 306, green 461, blue 616, left-most pixel column at x=0.
Juno rotation period=80.935 interframe delays, and roughly estimted to 30.063 sec, rotation phase of CCD around JunoCam's optical axis = 0.00153 radians relative to orthonormal frame defined by Juno's spin. Rotational angle around JunoCam's x-axis has been adjusted manually.
Geometry shift due to TDI has been neglected, as well as velocity aberration.
Assumed Jupiter shape model: 1 bar IAU Jupiter Maclaurin sheroid, rotating according to L3 longitude system.
Linear radiometric weights applied to decompanded colors: 0.88 for red, 1.0 for green, 2.17 for blue. Color interpolation is bilinear. Effects of global flat field and stray light have been neglected.
Resolution is 30 degrees per degree in an equdistant cylindrical system centered to the camera at image stop time, with an axis parallel to Juno's spin axis. The rendered field of view is 60x180 degrees.

Sorry about that. I've changed it to a smaller version.

I was pleasantly surprised that implementing the flat field correction cleared up the repetitive artifacts without need for any additional masking and filling. (For TDI=1 images)

FWIW, my current flat field derived color balance is {0.838223, 1., 2.37871}

Should be "Resolution is 30 pixels per degree..."

Gerald's sequence & details...

A couple of details using my updated pipeline

https://github.com/kmgill/cassini_processing



PJ15-19


PJ15-20


PJ15-21


PJ15-22


PJ15-23


Map projected combination of 22 && 23

Some of the approach images of the North Polar area


PJ15-11



PJ15-15



PJ15-17



PJ15-18

Here is a link to reprojected images of PJ15, part 2.

I'm currently busy with geometrical calibration for my EPSC talk. I may upload the images to missionjuno and UMSF, after I'll have completed a calibration milestone I've almost accomplished.

With the flat field, I've conservative in oder to be ready for DCT compressed images, for TDI >> 1, and for energetic particle hits. But with the Huffman encoded images, it may be worth to work directly with the flatfield.

People wanted the images redder. So, I changed to red=0.88. But this may be connected to the strong enhancement I'm usually applying, and the physically non-straightforward human color perception of poorly satutated colors. I might investigate the colors more thoroughly at some future point.

Thanks for the correction of my text! Late at night, and working on several tasks in parallel, I've to live with making errors.

Now, the enhanced reprojections of part 2 of PJ15 I've linked to, above:

PJ15, #24:
Click to view attachment

#25:
Click to view attachment

#26 and #27:
Click to view attachment Click to view attachment

#28 and #29:
Click to view attachment Click to view attachment

#30 and #31:
Click to view attachment Click to view attachment

#32, and #33:
Click to view attachment Click to view attachment

Btw., I've completed the calibration milestone. But the above images have been rendered with the parameters I've provided after the posts with the part 1 uploads.

Results of the new calibration iteration step are: focal length 1482.29 pixels, optical center (809.09; 609.20), rotation around optical axis 0.00145 radians, K1=-5.75E-8, K2=2.73E-14. The origin (0; 0) of the applied frame is the lower left corner pixel of the CCD (the side with the red filter).
But I think, that it's still possible, that the corners of the CCD are off up to 10 pixels with this approximation. However, that's outside the readout regions. Nevertheless, I'll try to narrow this further down with higher-order approximations.

Gif of Gerald's Perijove 15 images:
https://i.imgur.com/okSM2pI.gifv

Is this the final tally of images for Perijove 15?

No, there are still several more to come.

[edit: as of 14:36 PDT all of PJ15 should be on missionjuno.]

Couple more shots from Perijove 15



PJ15-35


PJ15-36


PJ15-37


PJ15-38


PJ15-39


PJ15-40


PJ15-41


PJ15-43


PJ15-58

And more



PJ15-24


PJ15-25


PJ15-26


PJ15-27


PJ15-28


PJ15-30


PJ15-31


PJ15-32


PJ15-33

Image PJ15_23, approximately true color/contrast:

Click to view attachmentClick to view attachmentClick to view attachment

And enhanced versions:

Click to view attachmentClick to view attachmentClick to view attachment

IMAGE_TIME = 2018-09-07T01:04:47.244
MISSION_PHASE_NAME = PERIJOVE 15
PRODUCT_ID = JNCE_2018250_15C00023_V01
SPACECRAFT_ALTITUDE = 6651.4
SPACECRAFT_NAME = JUNO
SUB_SPACECRAFT_LATITUDE = 38.1981
SUB_SPACECRAFT_LONGITUDE = 330.5426
TITLE = PJ15 North North Temperate Belt
Resolution at nadir: ~4.5 km/pixel

In the images above the resolution is highest (~4.5 km/pixel) within the brown barge near the right edge of images where a large part of the barge is visible.

These are heartbreakingly beautiful.

Extraordinary perspective and views, Kevin.

I was curious for a processing pipeline comparison, so I did up the same angles as posted by Bjorn on Sunday.

As typically output by my pipeline:
Click to view attachmentClick to view attachmentClick to view attachment


And enhanced using my normal methods (Photoshop + Lightroom)
Click to view attachmentClick to view attachmentClick to view attachment

It's great to see these pipeline comparisons!

And your pipeline Kevin is going from strength to strength, beautiful results with every update. *chef's kiss*

It should be good news for both of us that the results of the two pipelines look overall similar. Something weird is happening near the lower right corner in one of the images though. Are you using ISIS3 here?

Yes, see that? I am using ISIS3 and a lot of the maps get that little notch. I tend to crop around it, though.

If it wouldn't be too much trouble, if you could put together an example of what steps you followed to create this image, it might help the ISIS3 group debug the problem. This seems like too big a glitch to be attributable just to our camera model being a little off.

The process is automated here: https://github.com/kmgill/cassini_processin...m/processing.py

It's pretty straight forward use of junocam2isis, spiceinit, trim (optionally), and cam2map. (the rest of this process doesn't go into the above images). The resulting maps look reasonable:

Click to view attachment

The problem is probably more related in how I'm doing the vertex and UV mapping in the render code (https://github.com/kmgill/cassini_processin...junocam_cube.py). It's not super exact and in need of some attention.

I had assumed you were doing that in ISIS3 as well but it probably can't do an off-center point perspective view (my software can't either.)

Thanks for sending the code, that was very helpful!

Here are some more PJ15 renditions:
- PJ15, part 3, reprojected and enhanced,
- maps with gamma=4 thereof,
- gamma=1 version,
- PJ15 departure drafts.

I'll upload some of the reprojections to missionjuno later today.

PJ15_38 [Swift/Doran]

Here JPG copies of some of the PJ15 images, part 3:
#35, #36
#37, #38
Click to view attachment Click to view attachment
Click to view attachment Click to view attachment

#39, #40,
and #41:
Click to view attachment Click to view attachment
Click to view attachment

Here's PJ15_32 (SEB South). First approximately true color/contrast versions:

Click to view attachmentClick to view attachmentClick to view attachment

And versions with enhanced color and contrast:

Click to view attachmentClick to view attachmentClick to view attachment

And a subset of the associated metadata:

IMAGE_TIME = 2018-09-07T01:26:44.262
MISSION_PHASE_NAME = PERIJOVE 15
PRODUCT_ID = JNCE_2018250_15C00032_V01
SPACECRAFT_ALTITUDE = 11949.2
SPACECRAFT_NAME = JUNO
SUB_SPACECRAFT_LATITUDE = -22.4733
SUB_SPACECRAFT_LONGITUDE = 352.7984
TITLE = PJ15 South Equatorial Belt south
Resolution at nadir: ~8 km/pixel

PJ15 HD flyby movie on YouTube.
Link to MP4 version, according scenes and zipped JPG stills:

Beautiful work as usual Gerald.

Image PJ15_43. The left version is an approximately true color/contrast image. In the version at right the effects of the varying illumination have been removed.

Click to view attachment

This image was processed with a new version of my JunoCam processing pipeline after I fixed an annoying bug that was usually not of significance but could occasionally result in some color misalignment near the limb. This had to be fixed during postprocessing - not having to fix this misalignment saves a considerable amount of time when processing the images where this problem would have occurred.

By JunoCam standards this is a fairly distant image as this subset from the metadata shows:

IMAGE_TIME = 2018-09-07T02:21:50.387
MISSION_PHASE_NAME = PERIJOVE 15
PRODUCT_ID = JNCE_2018250_15C00043_V01
SPACECRAFT_ALTITUDE = 103371.3 km
SPACECRAFT_NAME = JUNO
SUB_SPACECRAFT_LATITUDE = -78.5043
SUB_SPACECRAFT_LONGITUDE = 66.4235
TITLE = PJ15 Southern timelapse sequence
Resolution at nadir: ~70 km/pixel

Amazing how far into the twilight-dark you get before it come apart. Very nice.

Image PJ15_22 in approximately true color/contrast and enhanced versions:

Click to view attachmentClick to view attachmentClick to view attachment

Click to view attachmentClick to view attachmentClick to view attachment

And a subset of the metadata:

IMAGE_TIME = 2018-09-07T01:01:47.358
MISSION_PHASE_NAME = PERIJOVE 15
PRODUCT_ID = JNCE_2018250_15C00022_V01
SPACECRAFT_ALTITUDE = 9184.6 km
SPACECRAFT_NAME = JUNO
SUB_SPACECRAFT_LATITUDE = 45.536
SUB_SPACECRAFT_LONGITUDE = 326.9532
TITLE = PJ15 Jet N4
Resolution at nadir: ~6.2 km/pixel

Image PJ15_30 in approximately true color/contrast and enhanced versions:

Click to view attachmentClick to view attachmentClick to view attachment
Click to view attachmentClick to view attachmentClick to view attachment

The second image has lots of interesting details. The enhanced version of it has been processed to enhance both large scale and small scale features by using processing parameters that differ for different parts of the image, depending on the contrast and the amount of small scale details in the original unenhanced image.

And a subset of the metadata:

IMAGE_TIME = 2018-09-07T01:20:45.235
MISSION_PHASE_NAME = PERIJOVE 15
PRODUCT_ID = JNCE_2018250_15C00030_V01
SPACECRAFT_ALTITUDE = 6252.9 km
SPACECRAFT_NAME = JUNO
SUB_SPACECRAFT_LATITUDE = -7.3210
SUB_SPACECRAFT_LONGITUDE = 347.0286
TITLE = PJ15 South Equatorial Belt
Resolution at nadir: ~4 km/pixel