Rovers are usually deplored to distant planets and asteroids and these locations could be tens of millions of miles away from the base station on Earth. Given the constraint that data cannot travel faster than the speed of light, instructions from Earth could take several minutes to get to the rover and the rover will therefore need to be autonomous at times. To achieve this autonomy, artificial intelligence could be used in consort with images of the rover's environment to enable safe operation. There is a need for pictures of both the front and rear of the rover to be taken, this is important because the rover can be made to move forwards or backwards without making a complete revolution. Two dimensional (2D) pictures will not be of great use here, because the depth of the scene will not be captured in 2D and this could lead to miscalculations that could result in the rover crashing into a rock or any other hazard.

Three dimensional (3D) pictures will therefore be required, and we make this possible by taking STEREO PICTURES from a pair of cameras. Taking stereo pictures involve taking 2D pictures of the same image from two cameras that have been aligned to mimic the way human eyes capture images. The cameras provide a black-and-white 120 degrees view of the terrain. They are separated by a distance determined by the closest image they will capture. They are angled inwards so that the intersection of imaginary rays projected from both cameras gives the depth that will act as a projection plane. Objects that appear before this plane will appear to be closer to the viewer and objects behind the plane appear to be further from the viewer. Using a number of techniques, the two images can then be viewed and the result is a 3D picture.

One way of viewing the pictures, will be to turn the two 2D pictures into an Anaglyph. An anaglyph is created by simply turning each 2D picture into a different colour and then superimposing both pictures. Viewing the resulting picture with corresponding colour of lens on each eye allows the viewer to see a 3D image.