Perseverance Route Map Options

[Phil Stooke]

Now we are down safely I am starting a map thread. Everything that happens during 2021 will be covered in my still-in-progress atlas, so I have to make the maps for that and everyone can follow as they evolve.

Let me know if any feature names start to show up.

Preliminary map even though not much has happened yet. But if you look closely you will see something, even now.

Phil

[kenny]

Lovely stuff, Phil.
Are we seeing evidence of the rock pavement being cleared of dust by the descent engines, to the NE and SW?
And conversely dark streaks towards the NW and SE ?
A very minor observation is that because the gridlines intersect in the middle of Perseverance (as they should), we can't clearly see your lovely little rover symbol!
Looking forward to seeing some white lines appearing soon....
Thanks again.

[Marz]

excellent! There's a white blob in the sand-filled crater 1 click SW of the rover. Seems strange for a large boulder to be there, so is that somehow part of the EDL?

[djellison]

Easy test - go find the location in this map. If it's in this map - it's pre-EDL - it's just natural.

You don't mean this by any chance? This is pre-landing.

Attached thumbnail(s)

 

[Andreas Plesch]

Following up on registering as accurately as possible a well suited rover downlook map image I thought it may be useful to easily look up coordinates and measure distances on the map. Here is a bare bones web app (takes a bit to load) to do that:

http://bit.ly/PercyMAP
with reduced texture for mobile
http://bit.ly/PercyMAPm

It is made using https://www.x3dom.org/ and a little javascript. One can zoom in (wheel or right mouse drag), pan (middle mouse drag) and re-center (double left click). The coordinates of the cross-hair are reported and left clicking adds the current position to a list, and calculates the distance from the last recorded point.

For example, it turns out that the wind-carved walrus boulder (harbour seal) is about 12m away from the rover. The used HiRISE DEM has nominally a 1m resolution but seems smoother than that in places, perhaps not picking up all boulders.

I may add a few additional features (controlling shading, continuous reporting of distance to rover) but nothing fancy to keep it really simple.

[edit] There is now a slider to control vertical exaggeration of the elevation model.

[Phil Stooke]

OK, first drive, first map. It's a close-up, not the overview provided by the first one I posted, which I will come back to as we move more.

This map uses HiRISE for geometric control, with descent images registered to that for more detail, and the first panorama projected out over that.

Phil

[PDP8E]

... and so it begins... surprise us!

[Phil Stooke]

OK, it did surprise me. We are outside my base map on sol 15, so here is a new map for this drive. That's the problem when you are mapping in real time! Fully prepared to have to switch map bases again tomorrow - that will be back to the first one I posted, most likely.

You may notice that my route is offset a few meters from the one on the interactive map on the mission website - they start a bit too close to the drift behind the rover at landing, and the whole path and each site retains that offset. My site locations are based on hazcam images.

Phil

[Andreas Plesch]

I had noticed that offset as well. It seems to me that somehow their rover position to base map registration is a little off.

Looking at the network traffic on https://mars.nasa.gov/mars2020/mission/where-is-the-rover/, it turns out that the traverse and waypoints are plotted with a geojson file.

traverse: https://mars.nasa.gov/mmgis-maps/M20/Layers...0_traverse.json
waypoints: https://mars.nasa.gov/mmgis-maps/M20/Layers..._waypoints.json

The geojson has the actual coordinates, and one can use it directly for mapping:



The traverse is given by small multiple segments, shown as the grey points. The red waypoints have additional rover orientation data, yaw, pitch and roll. yaw is shown on the map.

https://mars.nasa.gov/maps/location/api/con...get?mission=M20 is the overall layer configuration. It has the urls of the geojson geometries.

It seem that the web site uses a terrestrial CRS (WGS84) for plotting. Perhaps the rover coordinates are actually for a Mars CRS. That may explain that slight offset depending on the base map projection.

[kymani76]

My take at the landing map and first drives. It uses combination of Rover Descent Camera footage combined with HiRise image of Perserverance on the ground. Descent stage exhaust marks are also clearly visible.
The geoJSON files are indeed projected to WGS84 datum, so you only need to change that to Mars projection (both sphere and MOLA work) for the track to align in the right place.
The new interface also offers geoJSON updates for Curiosity's position.

[kymani76]

Another take at the Perseverance's landing map, with all the hardware location marked. Blue line is EDL track from SPICE kernels (looks like a pre-landing solution).
It's curious that the rover landed less than km from 2018 landing ellipse center.

[kymani76]

It's already time to update...sol 16

[Andreas Plesch]

And here is my take:



I am assigning CRS ESRI:104971 - Mars_2000_(Sphere) - Geographic to the geojson coordinates, and use the equirectangular projection for plotting which is native to the HiRISE mosaic.

The arrows show the yaw orientation at the way points.

The little circles along the path are where positions are actually provided in the traverse geojson.

We are getting outside the range of the EDL Rover Downlook imagery. Perhaps it is worth looking through those again to see if there is one covering the new terrain in higher detail than the 25cm HiRISE mosaic.

[Phil Stooke]

Map updated to sol 16. I hope for some feature names to add to it, eventually. I will have to re-think my earlier maps a bit. The sol 16 site is just a tiny bit outside the edge of the last map so I can probably adjust it.

It's great to see these alternative views of the traverse here.

Phil

[Julius]

Dont know if its the correct site to post this but looking at the orbital view, the rock pavement next to where the rover is sitting right now looks morelike sedimentary to me rather than a lava field and yet i am no geologist so who knows.

[Andreas Plesch]

I did find a slightly earlier frame of the Rover Lookdown camera which has a bit of the current location and is higher resolution than the HiRISE mosaic. It is a bit blurry because the Descent Stage had not fully stabilized and is from a more oblique perspective. But the registration went pretty well. Here is how the map looks with that frame (and higher contrast):



The width of the traverse path is 0.5 m.

The boulders are imaged quite a bit better than in the HiRISE mosaic.

[Phil Stooke]

I revised my map from earlier to include the sol 16 drive.

Phil

[Phil Stooke]

Iteratively building these maps. I have added the two SuperCam target names now, one on each map.

Phil

[MarkL]

Thanks as always Phil. This is a great resource.

[markril]

For those of you who might want to navigate this trip with Google Mars/Earth, as well, I've prepared some overlays using the large basemaps created by the USGS that were mentioned earlier in this thread.

The overview map is 5322x5040 with 6m/pixel resolution:



The larger 6.9GB map was broken into manageable 4000x4000 tiles with 0.25m/pixel resolution:



You can see some overlays of the various HiRISE and rover downcam images.

For the rover track, I cobbled together a script that scrapes the "Where is Perseverance?" web page. Right now I have to run the script manually, but plan to automate it once I can make it more robust.

You can download a zip file (55MB) with everything or download files individually at the site.

Additional map tiles will be added to the site soon to accommodate the planned route.

The elevation data currently used by GE is low resolution, so only the largest features like the crater rim render well with everything else looking kind of flat.

One thing I like about GE is that it provides an easy way to geolocate the interesting images presented in this forum for future reference.

Mark

[markril]

... and to continue the discussion on registration of the json track data from the "Where is Perseverance?" page with the USGS basemap, here's the post-landing HiRISE image layered on top of the Google Earth tiles:



This overlay and the downcam overlays (below) were manually registered (to the best of my ability and patience) with the Google Earth tiles derived from the USGS basemaps. These tiles were extracted from the original equirectangular basemap with a program using the supplied metadata (so no human inaccuracies).



So, yeah, there seems to be general agreement with the other maps here...

Mark

[Phil Stooke]

This is the sol 20 drive. The location should be good, the path is an estimate.

Phil

[Andreas Plesch]

Here is my best estimate of the current location, at sol 20 after the drive:



The rectangle is the approximate foot print of the rover, to scale.

This is mostly based on analysis of this Navcam image:



Starting from the dark, small, almost boulder free crater in the middle ground (apparently a word), I believe one can identify sets of boulders to track back to the rover. I guess we will see soon enough.

The crater seems really smooth and may be a good arena for heli testing. There are a few ripples, so perhaps the rover would not venture inside.

[kymani76]

Sol 20 update with newly available raw EDL images.

I would not recommend GE for geo-referencing of images as it only allows 4 (corner) control points.

[Andreas Plesch]

My estimate was a few meters off. With the latest geojson:



The width of the path corresponds to the width of the rover (ca. 2.7m).

https://mars.nasa.gov/mmgis-maps/M20/Layers..._waypoints.json now also has a few more fields: distance, drive and elevation among them.

In the helicopter introduction last night, they mentioned that its range could be about 100m x 30m, after initial shorter tests.

[markril]

For the Google Mars track, there's now an automatic update ability available for the client side. What this means is that Google Mars will check for an update every 6 hours (while GM is open) or you can manually refresh the track. This will download the track from my server. However, the track on my server is still only updated manually by me running a script so there will be some (human induced) latency until I fully automate the back-end (which will be soon).

Download & open track-updater.kmz and move it to the "mars2020" folder on the GM side panel. You can delete the old "track sols 2-16" folder. The new "track-updater" folder will automatically refresh in the future. To manually refresh, there's a refresh menu item when right-clicking track-updater. If you want, you can change the refresh interval or disable it by going to the Properties>Refresh tab.

Here's a look at the current track:



Note: While all the waypoints and traverses are displayed correctly, they are listed out of order on the side panel. This will be fixed in the next automatic refresh now that I've studied the json data a little closer.

Mark

[Phil Stooke]

The images for the last couple of days have not been helpful for making panoramas so it's taken a while to get to this map. Update to sol 23 - a drive back along the tracks away from the belly pan.

Phil




I register my maps to the HiRISE orthophotos, so they share projection details with those images. There may be small mis-matches between the various maps appearing in this thread but they are not worth worrying about, they all tell essentially the same story.

[Andreas Plesch]

Here is a map using the latest geojson, with registered EDL downlook imagery. The width of the path is 2.7m, the black circles are 1m in diameter. The faint arrows show yaw as reported in the way points. The projection is orthographic, centered at the landing ellipse (77.4298 E, 18.4663 N).

[nprev]

Some members have observed, and I agree, that this thread has drifted a bit from its intent. For our new members, Phil's map threads have generally served as navigation/situational awareness reference resources on UMSF for every rover since Spirit to help us all follow the journeys in those ways. Naturally, some questions and observations occasionally pop up in them but this has turned into a full-blown tech discussion.

I was gonna create a new Mars Cartography thread, but lo and behold it turns out that we have one already conveniently called Mars Cartography. 15 posts have been moved there, and I encourage all to please continue that most interesting and informative discussion there. Thanks!

[Phil Stooke]

Good idea.

I have adjusted the base of my intermediate scale map to include the Belly Pan location, so here is an update to sol 23.

Phil

[Phil Stooke]

This map shows feature names around the landing area mentioned at LPSC today. There are some diacriticals which I have not figured out yet, but which will be added later.

Phil

[kymani76]

sol 23 update
Only Rover Descent Camera images are used.

[Phil Stooke]

This is my previous map with the appropriate characters added.

Phil

[Phil Stooke]

Sol 30 update:



Phil

[kymani76]

map update

[Phil Stooke]

Using Hazcams I have plotted the path to sol 31.

Phil

Attached thumbnail(s)

 

[Phil Stooke]

The partial panorama in the early drives thread gives the path and location for sol 32.

Phil

[Art Martin]

Phil, I took your route map and superimposed a photo from JPL of the flight zone of Ingenuity. You can see they got very close to the drop off point with the latest drive.

location_sol_32 + flight zone

[Phil Stooke]

Thanks! Very useful.

Phil

[Andreas Plesch]

Another map based on the NASA geojson with simulated wheel tracks, 2.7m apart, pills from the traverse data, yaw arrows at way points, the airfield target and the overlook position:



And a full map with more context:

[kymani76]

Sol 32 update.

[Phil Stooke]

We have just 5 images from sol 33 down from Perseverance as I write this, but with one Navcam and one rear Hazcam there is enough information to approximate the drive:



Phil

[Andreas Plesch]

sol 33 update: The flight zone is 30m wide in this map.



full map

[Phil Stooke]

Quick update for the drive to the airfield on sol 34.

Phil

[pospa]

Quick update for the drive to the airfield on sol 34.

Phil, could you please (at least for the Ingenuity test period) mark the Helipad (3x3 m) and Airfield (10x10 m) borders into your map for our better orientation?
Thanks

[kymani76]

Sol 34 map update. Colored using merged RGB HiRise image.



Airfield closeup.

[kymani76]

https://arkeomapia.si/3d/mars2020/index.html
3D version of the map showing the flying zone. Not perfect, but you can pan around freely.

[MarkL]

[quote name='Andreas Plesch' date='Mar 25 2021, 10:43 PM' post='251192']
sol 33 update: The flight zone is 30m wide in this map.

So nice. Thank you!

How did you place the tracks? Visual references or are they just parallel to the path?

[Andreas Plesch]

How did you place the tracks? Visual references or are they just parallel to the path?

Glad that it is useful. Making the maps helped me understand the area better. The simulated tracks are just parallel to the traverse path from the geojson data, with an offset to indicate the 2.7m width of the rover. I think the plotted track marks are 40cm wide. I could not figure out if the cited 2.7m width refers to just the body of rover, or if it includes the wheels sticking out on the sides. I am using QGIS which makes makes it possible to do all that pretty quickly and accurately. It would be nice to simulate the rotational turns but I am not sure if there is enough data available (other than the imagery) and also do not have an idea how at this point.

[MarkL]

Thanks Andreas. Keep up the great work!

[Andreas Plesch]

sol34 update: larger airfield to avoid crowding, simulated tracks 2.7m wide from outer edge to outer edge (I think more correct), track marks 30cm wide. Flight zone 30m wide.

full map:

[Phil Stooke]

A special map for the area where the belly pan and debris shield were dropped. This is another use for the circular panoramas - they can be warped a bit to fit a HiRISE (or descent camera) image base.

Phil

[markril]

Here's the latest addition to the Google Mars project I've been working on:




Basicly, it's an image browser for Google Earth that organizes available raw images according to their geographical coordinates. The coordinates are derived from the image metadata. The raw images are generally released prior to updates to the "Where is Perseverance?" page so this gives advance knowledge of the rover position as determined by the navigation team. In any case, you can just view the images and make your own determination as I like to do.




The following web page contains further information and downloads:

http://vps78674.vps.ovh.ca/Mars2020/GoogleMars/

Mark

[Phil Stooke]

Interesting! Thanks.
Here is a map update for sol 43.
Phil

[Phil Stooke]

I have added recent drives and the helicopter deployment to my close-up map.
Phil

[htspace]

Here's the latest addition to the Google Mars project I've been working on:




Basicly, it's an image browser for Google Earth that organizes available raw images according to their geographical coordinates. The coordinates are derived from the image metadata. The raw images are generally released prior to updates to the "Where is Perseverance?" page so this gives advance knowledge of the rover position as determined by the navigation team. In any case, you can just view the images and make your own determination as I like to do.




The following web page contains further information and downloads:

http://vps78674.vps.ovh.ca/Mars2020/GoogleMars/

Mark

That's great, thank you!

[kymani76]

Sol 44 3D map view roughly matching today's self portrait.

[Phil Stooke]

We had a good drive on sol 47, probably about as long as the longest one to date at 35-40 m or so - I don't know the path at the start of the drive yet. We don't have a full panorama yet but the location should be fairly good.

Phil

[markril]

I mentioned in my last post that the raw-image metadata can be used to plot the position of the rover prior to the official geojson track being available and here's an example for sol 47. Image locations are the blue squares:



These coordinates appear to be based on the rover's dead reckoning abilities. Once the official track comes out with the latest localization, the image locations can be refined by distributing the error along the last traverse. This is all discussed (in great detail) in this oft-mentioned document.

One other interesting thing to note is that a stereo pair of Navcam images are generally downlinked 7-8 meters prior to the spot where the rover stops. Presumably this is an easy way to get an idea of what exactly is under the rover.

[htspace]

We had a good drive on sol 47, probably about as long as the longest one to date at 35-40 m or so - I don't know the path at the start of the drive yet. We don't have a full panorama yet but the location should be fairly good.

Phil

Perseverance has moved 38.43 meters on Sol 47. In total, it has already covered a distance of 229.74 meters.

[Phil Stooke]

Another drive on sol 48, getting closer to the overlook.
Phil

[Phil Stooke]

Sol 49 - closer to the overlook, but not yet at the exact point denoted earlier. I don't suppose they really have to worry about being exactly on it. This location comes from the Hazcam projection I just posted in the Early Drives thread.

Phil

[kymani76]

Sol 52 update. Flight set up (almost?) complete.

[Phil Stooke]

And my map for sol 52.
Phil

[Andreas Plesch]

I had been away from my qgis setup. Here my update for sol 59, with new labels:

and the full map:

[Andreas Plesch]

I tried my best to register the high resolution helicopter overhead perspective, using about 40 control points:



The helicopter lens seems to be wide angle, fish-eye, with heavy distortion away from the center. So I used rubber sheeting ( thin plate spline ) to generate a map:



Since I did not use the tracks for referencing, the good match between the actual and simulated tacks shows that referencing of the rocks went well, and that the base map was also well registered.

[Andreas Plesch]

I registered the second oblique helicopter view along the flight zone, clipping the highly distorted, more distant view. Here is a composite map of the flight zone which Ingenuity traversed to today, probably for about 50m.

[Phil Stooke]

A short drive which I was not expecting on sol 65, pointed out by Paul in the other thread... Using Navcams I find a location hastily mapped here. I will fix up the base later.

Phil

[Andreas Plesch]

For what it's worth I registered the third helicopter perspective (to the west) to the map:



Here the larger flight zone:

[MarkL]

For what it's worth

I'd say its worth a lot Andreas. Thanks.

[Andreas Plesch]

Thanks. Since I used rocks and pavement patterns to register the images to the EDL basemap, it was an interesting exercise to get to know a large number of these features more closely in the vicinity of the airfield, and some larger ones in the distance. With real photogrammetry and knowledge of the lens and camera orientation one could do a better job but this is pretty good on a cm scale.

[Andreas Plesch]

Updated maps to sol66:



full map:

[Antdoghalo]

Just to think, we are mapping the first controlled flight paths on Mars!

[Phil Stooke]

I will have a map of flight paths eventually! For now an update after a short drive on sol 66, well behind Andreas in getting to this. I am reverting to a smaller scale map until I have a better sense of where we are going.
The offset between real positions and the JPL map continues - it's about 1 rover length. The maps on here are all corrected.
Phil

[Andreas Plesch]

A mosaic of the four Ingenuity images, registered to a base map:



I used Thomas' straightened version (thanks !) hoping that it would behave favorably in the distance but I had to crop anyways. (In fact, I think it made interpolating between control points perhaps less robust than the original would have.)

[Phil Stooke]

Good stuff, Andreas!
Here is my more detailed map updated to sol 69. I will add descent images to the lower part soon.
Phil

[markril]

Google Mars for sol 72:



Position is dead reckoning but looks within a meter or two.

[Phil Stooke]

Good location! Here's my version.
Phil

[Andreas Plesch]

another take:

[Phil Stooke]

A short drive on sol 73 crosses a sandy patch as shown at small scale here. The sites indicated by squares are where arm instruments were apparently used, though without daily updates I don't know much.

Phil

[Andreas Plesch]

A sol73 map using the updated geojson for the traverse:



The orange lines bracket the destination of flight FOUR, with the circle denoting a rough best guess.

[Phil Stooke]

I'm still trying to decide what map coverage Ingenuity will get, but here is a first attempt covering the first 3 flights plus the landed location after Flight 4. I show the approximate coverage by Ingenuity's forward-facing camera based on mapping by Andreas and a tentative additional image, not yet released, looking north on flight 3.

Phil

[Andreas Plesch]

I registered the flight map released today and digitized the paths for flights 3 and 4:



There are also outlines of the mapped DEMs along the paths. I am relieved that my estimate (orange) of the path and destination (circle) turned out accurate enough. In fact, I would have expected that the 3d scan would be centered along the flight path but it is shown a bit off-center in the released map. Perhaps it will be worth to map out the path of Ingenuity's shadow from the released flight 4 sequence.

Friday's flight 5 apparently will track flight 4 (as much as possible) and so should end up in the circled area.

[Phil Stooke]

Great work, Andreas. Here is a map for Ingenuity's flight 4 - the next flight will be added to this later. I have deviated a bit from Andreas's map by showing the return path leading to the actual landing point. Ingenuity landed east of its original position after flight 4, and of course the fifth flight will also depart from there.

Phil

[kymani76]

And my version....

[Thorsten Denk]

How did you pinpoint the return point with centimeter-precision ("131.46 m")?

Thorsten

[kymani76]

It's a GIS derived value from the midpoint to the landing point. There are of course uncertainties in Ingenuity's and midway point position.
Should have really rounded up the number.

After some thought I think they moved the azimuth of the south flight zone area slightly to the SE direction in order to fit the descent camera coverage.

[kymani76]

Sol 77 / flight 5 update. Landing location estimated from Mastcam imagery. The official distance reported is 129m, but I get only 122.5m from the take-off and landing postions.

[Andreas Plesch]

https://mars.nasa.gov/resources/25900/nasas...n-new-airfield/

has a short video which includes this image of 'about where we landed':



This looks to me very similar to the flight 4 destination image so we may have ended up there.

[Apparently, the color camera on the helicopter is called the Return To Earth Cam, for reasons only known to the team].

[Phil Stooke]

This is a Mastcam-Z image pre-landing, roughly stretched to help estimate the landing location. By comparing with post-landing images I have added a likely landing point. It is a bit further out than on Kymani76's map, closer to the end of the previous flight. The 2 big rocks just above the midpoint of my arrow are visible in the descent image basemap.
Phil

[Phil Stooke]

Sol 84 - a drive to the west (a bit north of west really), possibly positioning for a nice smooth drive south to get closer to Ingenuity.
Phil

[Andreas Plesch]

sol 84 update from the updated waypoint and traverse geojson, and with a rough estimate of Ingenuity at its new airfield:

[Phil Stooke]

Just a short drive on sol 86, to put a knobbly rock in reach. We are back in the area of my more detailed map.

Phil

[Andreas Plesch]

Based on the flight 6 announcement and the high resolution color image from Ingenuity probably from when hovering high above its landing sight I made maps of the imaged area and of how flight 6 could look like:



It gives me confidence that the flight 5 landing location is not far off.

There is potential to improve the mapping by adding more control points. Hopefully I will find time to do that since it would improve on the HiRISE mapping.

The field of small, triangular dunes, and the east-west striping on the ground are among the features which could not be detected on the HiRISE images.

This is just a visualization of a potential flight path based on the description in the announcement. If I did that correctly, t works out for landing on the rugged crater rim (!).



The idea that drones could explore places inside craters not accessible by rover may be tested soon.

[Phil Stooke]

Sol 91 - a shortish drive to the south, maybe 25 m. The location is based on a partial panorama.

Phil

[Station]

Sol 91 - a shortish drive to the south, maybe 25 m. The location is based on a partial panorama.

Phil

Hi Phil,

Is this white rectangle (sol 76 - Airfield "B") the last landing location of Ingenuity? Is the Field C next possible target for little helicopter?

[Phil Stooke]

The symbol is the approximate location of the helicopter landing after flight 5 - still hoping for a new picture after the latest drive to nail down the position a bit better. Field C is the (approximate) next landing site.

Here is another version of the map.



Phil

[Andreas Plesch]

A larger map with the json traverse to the current site, the helicopter aerial image at the end of flight 5, now more carefully georeferenced, and a tentative flight 6 path, based on the narrative provided in the update news item. The current location of Ingenuity on the map may well be a little off.

[Phil Stooke]

This is my best guess at the moment for flight 6. I'm still not certain I have the end of flight 5 location correct. Flight 6 runs roughly SW for about 150 m, and where it stops it takes the aerial view out over the landscape, the one which has just been released. Then it runs south for about 20 m, and that's where the video starts. From there it runs NE and then turns a bit and moves north before landing. That landing point is right on Andreas's vector. Subject to modification!!! (as always).

Phil

[Andreas Plesch]

This is my best guess at the moment for flight 6. I'm still not certain I have the end of flight 5 location correct. Flight 6 runs roughly SW for about 150 m, and where it stops it takes the aerial view out over the landscape, the one which has just been released. Then it runs south for about 20 m, and that's where the video starts. From there it runs NE and then turns a bit and moves north before landing. That landing point is right on Andreas's vector. Subject to modification!!! (as always).

Phil

Thanks for giving flight 6 a try. I get the same spot for the location of the aerial view, and it is likely that the small spidery dunes in this area are the same which are shown in the video. Running south, and then turning sharply back to the NE (as planned) would also pass by the small crater as seen in the video. But I am not sure I can see the required sharp turn by a counterclockwise 135 degree from S to NE in the video. There is a turn but it looks more like 45 degrees. I will need fresh eyes to look again at the map, the panorama, and the video.