The HiRise camera on the Mars Reconnaissance Orbiter snapped a beautiful photo of Perseverance falling through the Martian atmosphere hanging from its parachute.

This photo can be found here on the University of Arizona's HiRISE page.

This answer explains that the HiRise camera is a push broom type; it scans a 1D-like sensor across a long swath of Mars, like older photocopy machines and fax machines. I suppose this means it couldn't shoot a sequence of shots.

HiRise Specifications: https://www.uahirise.org/specs/

The image page says:

HiRISE was approximately 435 miles (700 kilometers) from Perseverance and traveling at about 6750 mile per hour (3 kilometers per second) at the time the image was taken. The extreme distance and high speeds of the two spacecraft were challenging conditions that required precise timing and for Mars Reconnaissance Orbiter to both pitch upward and roll hard to the left so that Perseverance was viewable by HiRISE at just the right moment.

Question: Why exactly did MRO have to "both pitch upward and roll hard to the left" to point exactly at Perseverance during the entry, descent and landing phase? What does that even mean?

Since the target object is tall and skinny, did it somehow have to maneuver to "pushbroom" horizontally so that the scan was across their short direction rather than bottom-to-top?

PIA24270: HiRISE Captured Perseverance During Descent to Mars

Source: PIA24270: HiRISE Captured Perseverance During Descent to Mars

  • $\begingroup$ possibly helpful: High-Performance Pushbroom Imagers for Planetary Missions Bergstrom and Dissley 2012 (Ball Aerospace & Technologies Corp( $\endgroup$
    – uhoh
    Commented Feb 20, 2021 at 2:39
  • 1
    $\begingroup$ Are you sure it wasn't just to make sure that the telescope was pointing at the lander? $\endgroup$
    – Phiteros
    Commented Feb 20, 2021 at 5:19
  • 2
    $\begingroup$ The MRO was pretty low in the sky during the landing. They probably typically image just below the orbiter under normal circumstances, so it had to maneuver to be pointing at Perseverance. Since it was also being used as a communications relay, they probably also didn't want the antenna to be pointing away from Earth. That would be my guess. $\endgroup$
    – Phiteros
    Commented Feb 20, 2021 at 5:33
  • 1
    $\begingroup$ If you substitute "yaw" for "roll" does it make more sense? My guess is they felt like avoiding the term for whatever reason, substituting it with a more public friendly "synonym" $\endgroup$ Commented Feb 21, 2021 at 6:28
  • 2
    $\begingroup$ They probably did mean pitch and roll. Yaw doesn't do much for a nominally down-looking spacecraft. $\endgroup$ Commented Feb 24, 2021 at 4:35

2 Answers 2


Why exactly did MRO have to "both pitch upward and roll hard to the left" to point exactly at Perseverance during the entry, descent and landing phase?

Perseverance was not directly underneath MRO at the time the images were taken. Perseverence was instead forward of and to the left of MRO's ground track. The fairly standard meaning of yaw, pitch, and roll for an orbiting spacecraft are

  • Yaw is a rotation about the vertical axis,
  • Pitch is about the across-track axis, and
  • Roll is about the along-track axis.

Yaw is the last thing one wants to do with a nominally down-looking push broom camera. Think of sweeping the leaves from your sidewalk with a push broom. Yaw (in spacecraft convention) is the equivalent of turning the broom sideways. Instead of sweeping a half meter swath of the sidewalk you're sweeping a swatch that's a few centimeters wide.

Pitch also isn't normally used, but it may be needed to see a dynamic event. For a static event, the spacecraft's travel will soon take the spacecraft over the spot in question. But for a dynamic event such as seeing parachute deployment from above, the spacecraft may need to pitch up or pitch down so as to have the event in view.

The HiRISE instrument has a very narrow field of view. This is where roll becomes important. If the camera is to see some object (dynamic or static) that is not directly underneath the satellite's path, the satellite will have to roll to bring that object into the camera's field of view.


A year plus later and with the enormous benefit of reconstructed data (SPICE kernels), the exact maneuvers of the spacecraft can be visualized.

The PDS Label for the MRO data product gives the following start & stop times:

/* All xxx_TIME values are referenced to UTC.                           */
    /* Time when the observation first started */
    MRO:OBSERVATION_START_TIME   = 2021-02-18T20:42:31.657
    /* Time of the first image line */
    START_TIME                   = 2021-02-18T20:42:32.368      <- START
    SPACECRAFT_CLOCK_START_COUNT = "1298148206:43887"
    /* Time of the last image line */
    STOP_TIME                    = 2021-02-18T20:43:02.881      <- STOP
    SPACECRAFT_CLOCK_STOP_COUNT  = "1298148237:11987"
    /* Time when this RDR product was created */
    PRODUCT_CREATION_TIME        = 2021-02-25T00:45:18

The high gain antenna (HGA) must maintain Earth pointing as for Mars 2020 (unlike MSL) MRO is the "primary telecom link from atmospheric entry to rover landing" [1]. MRO & MAVEN are the primary links, but MRO achieves near real-time transmission in a "pseudo bent pipe" mode. MAVEN records open loop and relays back to Earth sometime later.

MRO's nominal attitude & spacecraft fixed reference frame is +X in direction of flight, +Z "down" (also the view direction for HiRISE), +Y completes the right hand cross product. The HiRISE CCDs scan in the X direction. This is taken from MRO's latest frames kernel (fk) & HiRISE instrument kernel (ik). Here is an animation I cooked up showing this nominal attitude just prior to Mars 2020's EDL:

(Personal work, data from NAIF, RGB -> XYZ, grey -> nadir)

It's tough to see the blue +Z line as it essentially overlaps the grey nadir line in MRO's nominal attitude. Both the solar arrays and the HGA on MRO have an inner and outer gimbal to maintain their own pointing (towards the Sun & Earth) distinct from the spacecraft bus attitude. These gimbals mean that any "off-nominal" MRO orientation during the EDL event is (almost certainly) solely for the purpose of taking the picture (including any potential momentum management schemes before/after the picture).

Contrary to at least my assumptions, MRO is out of nominal attitude for nearly half of its orbit and orienting for the HiRISE observation takes several minutes:

(Personal work, data from NAIF, RGB -> XYZ, grey -> nadir, yellow dashed -> towards M2020)

Here are two plots of the orientation angles during the sequence:

rotating frame angles inertial frame angles

(Personal work, data from NAIF)

In the rotating frame, left, the nominal attitude is seen as yaw, pitch, roll angles of ~0. We see MRO pitch up, roll left (negative roll), and yaw right (positive yaw). The spacecraft "over pitches" and actually pitches down during imaging, though still "up" from nominal attitude. During the imaging, the spacecraft brings the yaw angle back to near zero, as @DavidHammen explains is important.

I will point out though that Mars 2020 is actually ahead and to the right of MRO when the image is taken as seen in this wider view animation:

(Personal work, data from NAIF, RGB -> XYZ, grey -> nadir)

The prior plots show some motion during the imaging, which can be seen in the image itself. These artifacts of the spacecraft's maneuvering in the full image can be seen at the start (bottom) of the image:

MRO M2020 EDL image

(NASA/JPL/University of Arizona)

See the "squiggliness" on the right edge.


  1. Abilleira, F. et al. "Mars 2020 mission design and navigation overview," (2019) (JPL TRS link)


Interestingly, the Perseverance rover (and descent stage) are almost not in the zoomed in image. In the full size image (18048 px by 62300 px!), the backshell and parachute are located at ~64% of the vertical (scan/pushbroom) dimension. This means that the backshell and parachute are "imaged" approximately 64% of the way through the observation timewise. When Mars 2020's inertial velocity is plotted over the observation period:

M2020 inertial velocity

(Personal work, data from NAIF, vertical line indicates ~64% of way through observation time)

We see that the vehicle is imaged almost exactly when the descent stage and rover separate from the backshell (see the smooth increase in velocity).

Though it would be cool to see a free flying Perseverance, I can't see anything that looks like a descent stage in the full resolution image:

full res close up

(NASA/JPL/University of Arizona)

  • 2
    $\begingroup$ @compi thanks! I had a lot of fun going through this $\endgroup$ Commented May 20, 2022 at 13:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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