With SpaceX planning to land humans on Mars as soon as 2026, we have never been closer to landing on another planet — it’s in just 5 years!
But what about the 8 billion humans left behind? We also want to explore!
Sadly, the internet lacks a high resolution, natural color Mars. So that’s what we built.
While Google’s Mars map uses low resolution images taken in the 70’s, Google Earth only low quality image blends, NASA’s site only provides us with a number of different maps to choose from, which patch together in a mosaic of different resolutions and non-matching colors.
That’s why we built a website presenting Mars in natural color, high resolution images.
And that’s why we opened our source code, allowing anyone to add features and better images to the site simply by pushing PRs to the repository.
The best of three cameras
To create the most detailed Mars imagery to date we developed a process merging different image layers into one consistent map. Let us walk you through the high-level approach below.
The Viking Images were taken from 1976 to 1980, by the Viking 1 and 2 orbiters. In 2009, NASA’s Ames Research Center in California merged 4600 Viking Orbiter images with a black and white Mars Digital Image Model. The result is a colorized global mosaic with the resolution of about 230 meters per pixel at the equator. This model is what you see in most Mars maps and images today.
AMES’ model takes the varying color and brightness of the Viking images into account and adjusts the parameters slightly. Overall the model might still be more red-ish than is natural to the planet’s surface, since the filters used in Viking’s image setup lean towards red.
We use the Viking model as color for our higher quality image of Mars.
Our next layer, MOLA HRSC Blended DEM, is a digital elevation model that combines data from the Mars Orbiter Laser Altimeter (MOLA) and the High-Resolution Stereo Camera (HRSC). Published in 2018 by the USGS Astrogeology Science Center, it has a resolution of 200m per pixel with an accuracy of 3 meters in height.
MOLA flew aboard the NASA Mars Global Surveyor spacecraft (MGS) and used lasers to deliver elevation data with a resolution of 463 meter per pixel. 600 million measurements taken between 1999 and 2001 are added to the model and combined with the data taken from HRSC.
HRSC has flown aboard the European Space Agency’s Mars Express (MEX) spacecraft since 2004. Unlike MOLA, HRSC delivers an elevation model by taking images at different angles, an approach that delivers data with a resolution of up to 2.3 meter per pixel.
We use this DEM as elevation data.
Our third image layer is delivered by the Context Imager (CTX) aboard the Mars Reconnaissance Orbiter (MRO), which reached Mars in 2006. It takes grayscale images with a resolution of up to 6 meter per pixel. To date CTX imagery covers already more than 99% of the red planet’s surface.
There is way more imagery of Mars, most of which we could not use in our model. However one project is especially worth mentioning: Besides CTX, the MRO also hosts the HiRISE camera. HiRISE is recording the highest-resolution images of Mars yet, with a stunning 30 centimeters per pixel. However, HiRISE imagery covers only 5% of Mars’ surface. Full coverage might only be reached after humans already touched ground.
Creating the best, global Mars map imagery
With the source layers set, we now began to research how to merge them into a desired Mars map. The basic idea is to merge the Viking images as color layer with the CTX images as high resolution layer and the HRSC elevation model to map the Z-axis. Sounds simple, but involved months long research as well as programming a batch process to calculate the end product. Let’s look into the exact process below.
The Viking images provide color to the final layer. These images were taken during the late ’70s, roughly 10 years after the moon landing. And within that context they are astonishingly good.
However, you might see why we would want to improve them: Extreme bright and dark areas due to over and under exposure, low detail and varying contrast, brightness and color.
We now reduce extreme exposure on both ends of the spectrum, then blur the image, before taking it as a color overlay on top of the CTX layer. This way, we ensure that just the color, but no insufficiently sharp details are inherited to the end product.
For the over/underexposed areas where no color is available, we supplement an average color.
The CTX imagery on the other hand comes in high resolution, but only in grayscale.
The area not covered by CTX is shown as a black area. Those areas are, luckily, covered by the Viking imagery.
We now increase contrast on the CTX image and adjust the Viking layer to a color more representative of natural Mars. We then underlay the Viking image to the current layer, so that it can be seen through CTX’ empty areas, and finally improve transitions to get rid of sharp edges.
The whole calculation took about 100 hours on an average 8 core 64 ram machine. Then another few days for building images for a map server. This server offers small images to the front-end, based on a user’s current position on the Mars globe.
You can see why we are happy to finally release after running multiple of those iterations to improve the final result.
The end product is a simple website with a full Mars globe, in nature-true color with details from Mars’ high-res images with the highest existing coverage.
Visit mars26.com , explore closer than ever before and leave a Tweet at your favorite spots!
Stay tuned for our bigger plans; A nifty surprise is coming soon.
News: twitter.com/marstwentysix
Community: discord.gg/hByPR9GcbU
Github: github.com/mars-2026/mars26_com
Thanks to the NASA’s Ames Research Center for their work on the Viking Mosaic, Caltech’s Murray Lab for their work on CTX imagery, USGS Astrogeology Science Center for their work on MOLA HRSC Blended DEM as well as DLR for their work on the HRSC, NASA and Goddard Space Flight Center for MOLA, the University of Arizona for their work on HiRISE, Malin Space Science Systems for their operation of CTX, NASA Jet Propulsion Laboratory for various work involved and all institutions involved in the PDS. We are truly standing on the shoulders of deep space giants.
* Today’s featured banner image shows the infamous Schiaparelli Crater, named after Italian astronomer Giovanni Schiaparelli. From Schiaparelli the Italian term “canali” originated, which mistranslated into ‘canals’ made its way into late 19th century science and finally science fiction.
The crater is actually Mark Watney’s destination in ‘The Martian’ and sadly, a couple hundred kilometers from the crash site of the Schiaparelli EDM lander.
In the picture you can actually spot darker layers of volcanic stone where the red sands have been blown away by strong winds.
