4
$\begingroup$

I'm curious about this picture taken from Osiris-Rex. There is a faint vertical line nearly the width of Earth going from the top to the bottom of the image. What is it? Does it have something to do with the camera, or is it some sort of leftovers from cleaning up the image, or something else?

Here is the image with "Shadows Enhanced": https://i.stack.imgur.com/W0scW.png


enter image description here

$\endgroup$
  • 7
    $\begingroup$ If you could point out where the line is, it may help. I don't see any line. $\endgroup$ – B.fox Jan 2 '18 at 23:06
  • $\begingroup$ You need to add a link to the source of the image. Necessary information includes which camera was used! I'll add it for you this time. $\endgroup$ – uhoh Jan 3 '18 at 2:19
  • $\begingroup$ @B.fox I've added a link to the same image, but with "shadows enhanced". Since the question is about an artifact of the image, I think the idea was to leave the original image unaltered. $\endgroup$ – uhoh Jan 3 '18 at 3:15
7
$\begingroup$

The line you are referencing is because the Earth is very bright on a "black" background. It happens as the light overpowers the image chip, similar to the Starburst pattern bright headlights make on dark nights.

$\endgroup$
7
$\begingroup$

The image was taked by the OCAMS instrument. OCAMS uses CCD (Charge Coupled Device) focal planes. The charges generated by the incident light in each frame are transferred down the columns one pixel at a time. If there is an especially bright object, it may affect the remainder of the column.

Note: OCAMS is a suite of cameras: PolyCam, MapCam, and SamCam. which have narrow, medium, and wide fields of view respectively. All use the same CCD focal plane design.

Charge Coupled Device (CCD): https://en.wikipedia.org/wiki/Charge-coupled_device

OSIRIS-Rex OCAMS Instrument: https://arxiv.org/ftp/arxiv/papers/1704/1704.04531.pdf
(Note that the figures are at the end of the .PDF).

$\endgroup$
  • 4
    $\begingroup$ It does somewhat extend upward, though. $\endgroup$ – Rikki-Tikki-Tavi Jan 3 '18 at 0:05
5
$\begingroup$

The streak is perfectly aligned to the vertical axis. Assuming that it hasn't been rotated, it is likely to be an artifact from imager chip, which may have saturated or overloaded - more charge/pixel than they can linearly process. However I noticed that the color of the streak is pure white, which is hard to understand unless it is caused specifically by the white clouds.

The image of the Earth is slightly saturated, and it can be seen that the bumps in the vertically collapsed image of the earth correlate well with the bumps in the vertically collapsed streak, so while somehow this is likely the cause, the particulars may lie in the processing.

It is important to note that while Orisis' target Bennu has a similar distance from the Sun as the Earth, it's albedo is only about 0.05. The camera is optimized for darker images, so those white clouds may simply be brighter than the design range of the whole system.

Your last link says:

On October 2, 2017, the MapCam instrument on OSIRIS-REx captured the data for a composite image (above) of the Earth and Moon. The spacecraft was approximately 5 million kilometers (3 million miles) from Earth at the time, about 13 times the distance between the Earth and Moon. (Click here to see the geometry of the shot.) Three images (different color wavelengths) were combined and color-corrected to make the composite, and the Moon was “stretched” (brightened) to make it more easily visible.

and according to Spaceflight 101:

MapCam is a medium-range camera. It searches for satellites and outgassing plumes around Bennu, maps the asteroid in color, and provides images to construct topographic maps. Its filter wheel and five-element lens system allows both panchromatic (clear) and wide-band spectral imaging in the blue, green, red, and infrared.

Read more technical information in OCAMS: The OSIRIS-REx Camera Suite

enter image description here

enter image description here

import numpy as np
import matplotlib.pyplot as plt

img = plt.imread('FWC1z.png')[...,:3]  # get rid of alpha

print img.shape

earth  =   img[244:284, 150:250]
streak = 2*img[400:800, 150:250].reshape(40, 10, 100, -1).sum(axis=1) # vert bin by 10

plt.figure()
plt.subplot(2, 2, 1)
plt.imshow(earth) 

plt.subplot(2, 2, 3)
for thing in earth.sum(axis=0).T:  # plot r, g, b separately
    plt.plot(thing)
plt.ylim(0, None)

plt.subplot(2, 2, 2)
plt.imshow(streak) 

plt.subplot(2, 2, 4)
for thing in streak.sum(axis=0).T:  # plot r, g, b separately
    plt.plot(thing)
plt.ylim(0, None)

plt.show()

earth_2 = earth[:, 20:80]

plt.figure()
plt.subplot(2, 1, 1)
plt.imshow(earth_2)

titles = "red", "green", "blue"
for i, (color, title) in enumerate(zip(np.rollaxis(earth_2, 2), titles)):
    plt.subplot(6, 1, i+1+3)
    print "i = ", i, "title = ", title
    for line in color:
        plt.plot(line)
    plt.ylim(0, 1.05)
    plt.text(3, 0.6, title, fontsize=16)
plt.show()
$\endgroup$
4
$\begingroup$

Largest original: https://www.asteroidmission.org/wp-content/uploads/2017/10/EGA_Plus_10_mapcam_Earth_Moon.png

Blooming is vertical while streaking is horizontal:

Blooming vs. Streaking

Some people use blooming and streaking interchangeably, while others use them as synonymous with each other. Some CCDs have anti-blooming gates which mostly inhibit the problem but can result in as little as 70% fill factor with resulting loss of sensitivity and well depth.

The charge capacity of an image sensor can be limited by either the individual photodiode characteristics (pixels) or the CCD itself, and is defined by the maximum amount of charge that the image sensor can collect and transfer while still maintaining all of its design performance specifications.

This capacity limit is termed the saturation charge level, and when this limit is reached, the pixel or CCD is described as being saturated. Exceeding the saturation level results in the generation of blooming artifacts in captured images. Blooming refers to the overflow of excess confined photo-generated charge from a photodiode well into adjacent structures when the maximum well charge capacity is exceeded.

Smear is generated directly or indirectly in the vertical shift registers (VCCD) in the interline transfer CCD (IL-CCD). The VCCD is a light-shielded area of the image sensor used to transfer the charge off of the sensor.

Crop of enhanced image:

Image enhanced and tone mapped on blue channel

Hamamatsu article on "CCD Saturation and Blooming".

Video: "What is the Blooming effect in CCD chips?".

Note that there is a band on the left side and numerous speckles which tends to indicate that a minimal amount of processing was done on the image; knocking down the blooming was likely the extent of what they wanted to accomplish.

Calculations of the effect in CCDs: https://www.pco.de/fileadmin/user_upload/pco-publications/pco_pub_20000107_smear_e.pdf

Here is an example and explanation of why the bloom is vertical:

CCD vs. CMOS Imagers

The CCD (although best in class -120 dB) shows the problem while it's not visible in CMOS (no vertical charge transport).

See also Apogee Instruments article on Blooming vs. Anti-blooming CCDs.

FSU's "Understanding Digital Imaging" webpage shows this image:

CCD internal structure

$\endgroup$
  • $\begingroup$ Blooming is vertical while streaking is horizontal: that would depend on the orientation in which the CCD is installed, wouldn't it? $\endgroup$ – Hobbes Jan 3 '18 at 16:04
  • $\begingroup$ @Hobbes - No. hamamatsu.magnet.fsu.edu/articles/ccdanatomy.html - Direction is referenced to the internal elements, not the physical orientation. The saturation is distributed horizontally or vertically across the surface of the chip, the effect produced is dependent upon which elements are effected (the saturation travels horizontally or vertically). It's not like a polarizer where you can turn it to a specific angle and then eliminate the effect, or be smart and turn it 45° then proclaim that it has neither horizontal nor vertical artifacts. Good question. $\endgroup$ – Rob Jan 3 '18 at 16:43
  • 1
    $\begingroup$ Excellent answer! Thanks for adding all of the additional links. This will be a helpful resource for the site. $\endgroup$ – uhoh Jan 4 '18 at 1:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.