The Face on Mars

© Copyright 1995 Malin Space Science Systems, Inc.


Table of Contents

Additional Information on "The Face On Mars"


In July, 1976, Viking Orbiter 1 was acquiring images of the Cydonia region of Mars as part of the search for potential landing sites for Viking Lander 2. On 25 July, 1976, it photographed a region of buttes and mesas along the escarpment that separates heavily cratered highlands to the south from low lying, relatively crater-free, lowland plains to the north. Among the hills was one that, to the Viking investigators scrutinizing the images for likely landing sites, resembled a face. Owing to the importance of the landing site search, and with a desire to provide the public with at least one familiar-looking landform amid the craters and exotic terrains found all over Mars, an image including the face-like hill was released as part of the Jet Propulsion Laboratory's public relations effort. The text of that release notes the face-like hill.

Subsequent to this release, some people have argued, mostly in the lay literature, that the face-like hill is artificially shaped. Although their argument has been expanded to a host of nearby features, none commands public interest like the "Face." This page will provide interested persons with both the raw Viking images, transformed to GIF format, and a brief tutorial (with examples) of image processing techniques applied to create "better looking" images.

The Raw Images

Table 1, below, lists all Viking Orbiter images of the "face," in order of decreasing resolution. Following the table are the images acquired at resolutions better than 400 meters/pixel. The icons show the "face" as seen in the raw data, but contrast enhanced for visibility. The actual images have not been processed other than conversion to GIF format. As a reminder, each raw Viking Orbiter image is 1204 samples wide by 1056 lines high.

Table 1: List of Viking Orbiter Images of the "Face on Mars"

*070A1343 m753A06588 m
*035A7247 m753A03596 m
*561A25163 m717A04723 m
*673B56226 m771A94735 m
673B54226 m257S41780 m
*753A34233 m220S38808 m
753A33233 m257S69821 m
753A05578 m220S37863 m
859A05582 m590A24889 m

*images included in this Web page

Raw Viking Orbiter Images






Label Information (Ancillary Data)

Processing the Face Images

The following images were subjected to image processing techniques to improve the visibility of features in the images. A step by step description of image processing techniques as applied to these images is also on-line. The processing applied includes bit-error correction, reseau removal, very slight brightness alteration, and projection to a standard map view (mercator projection) with north at the top. Contrast/Brightness enhancements and image sharpening were not performed as these would create images that would differ depending on the nature of the monitor on which they were being viewed. Examples of this type of processing are given elsewhere. Map projection resamples the raw image format, and these images are larger than the raw data. The following table lists the dimensions of these images:

Table 2: Dimensions of Mercator Projected Images of "The Face on Mars"


Map Projected Images
035a72m 070a13m
561a25m 753a34m


Derivation of a model of the relief of an object from the way it reflects light seems reasonably straightforward: slopes facing a light source appear brighter than those facing away. However, in practice this is a difficult problem to solve, and the results are often non-unique. Based on the computer vision literature, MSSS scientists have developed a "shape-from-shading" technique that can be applied to planetary images. The following images are the vertically viewed height field (exaggerated to fill an 8-bit, 255-level range) and a stereoscopic view of the original image draped over the height field in both stereopair and anaglyph form.

Topography (GIF = 38 KBytes)

Stereopair (GIF = 52 KBytes)

Anaglyph (GIF = 19 KBytes)

Changes of Appearance with Illumination Direction

Using the height field, it is possible to view, at the admittedly lower spatial resolution of that field, what the "Face" looks like under different illumination conditions, in particular from illumination at different hour angles. Since the "Face" is above the martian "Tropic of Capricorn" (+25 deg), illumination in reality will always be from a southernly direction, but in these simulations, illumination can come from other directions as well.

Illumination from southwest

Illumination from south

Illumination from northeast

Illumination from north

MPEG of 360 deg rotation of illumination direction

Changes of Appearance with Viewing Direction

Using the height field, it is also possible to view the "Face" from different look directions, by mapping the Viking Orbiter image directly on the the topography and then placing the "viewer" at different locations.

View from southwest

View from south

View from northeast

View from north

MPEG of 360 rotation of viewing direction




Calvin J. Hamilton