# Mapping the moon: how to use the Lunar Reconnaissance Orbiter (LRO) image data?

what's the meaning of Scaled pixel width and scaled pixel height of LROC EDR image obtained by LROC image search tool?

– uhoh
Nov 6, 2016 at 2:35
• I'm pretty sure (but not 100% sure) those represent the corresponding approximate scale of the image in meters per pixels. They need to be calculated carefully considering the exact position in the orbit relative to the moon, and the angle that the camera is pointing, and the actual scale will vary within the image due to geometry and topography. I believe that the link to that data is wms.lroc.asu.edu/lroc/view_lroc/LRO-L-LROC-2-EDR-V1.0/…
– uhoh
Nov 6, 2016 at 3:28
• Page 7 show that by the time this image was taken 2012-01-15 LRO was in a rouhgly 185x40 km orbit: hou.usra.edu/meetings/leag2014/presentations/keller.pdf
– uhoh
Nov 6, 2016 at 3:34
• @uhoh yeah that exactly is my data. Thank you very much. I wanna find the actual length of an object in this image. I'm getting the pixel length in matlab. so is it correct to use this scale appropriately to find the real length in meters? Nov 6, 2016 at 8:16
• I left a comment but not an answer because I am not 100% sure. If you are doing something just for fun, then you can find an image with a known crater, or feature, try this and compare to a known good map. Make sure to check x and y independently. However, if you want to be sure and correct, then we'll have to keep digging, and at the same time someone may come along and post a real answer. Remember, the moon is a 3D topographical surface and the images are taken from different angles, they are not all "top-down".
– uhoh
Nov 6, 2016 at 8:29

tl;dr Those numbers are meters per pixel. But you have to be careful because they are scans along the spacecraft orbit with the moon rotating underneath, so x and y are skewed.

This answer will get you started. The LRO is an amazing spacecraft with a wide variety of instruments to map the moon's surface by a number of characteristics as well as simultaneous laser ranging information for LRO-Earth and LRO-Moon paths improving lunar gravity models. There is a very helpful one page overview in The Lunar Reconnaissance Orbiter and the New Moon: Mission Highlights and Two More Years of Science From Lunar Orbit! by project scientist John Keller. You are interested in the LROC/NAC (LRO Camera / Narrow-Angle Cameras)

The LRO Cameras are described in detail in Lunar Reconnaissance Orbiter Camera (LROC) Instrument Overview (Robinson et al. 2010, Space Sci Rev (2010) 150: 81–124 DOI 10.1007/s11214-010-9634-2):

There are two essentially identical narrow angle cameras, named "L" and "R" in the data arcives. Each camera uses a roughly 5000 pixel by 1 pixel CDD, and the LRO's orbits in the "X" direction continiously scans like a FAX machine or flatbed scanner. The FOV of the cameras slightly overlap in the middle, but extend to the left and right of the centerlin thus the "L" and "R" designations in the image data archives as well.

Much of the data is taken in a roughly polar orbit, including the data in January 2012 you've mentioned, so the ~5000 pixels of a given camera in this case are nearly parallel to lines of lunar latitude.

Upper right latitude    42.08
Upper right longitude  340.64
Lower right latitude    44.76
Lower right longitude  340.7
Lower left  latitude    44.76
Lower left  longitude  340.33
Upper left  latitude    42.09
Upper left  longitude  340.28


From that we can plot the approximate border of the image:

The skewing is due to some combination of the inclination orbit rotation of the moon under the spacecraft while it scans. Remember to consider the skewing when trying to use these images!

The width is about $340.7-340.33\approx0.37$ degrees in latitude. The data below show that the camera is looking nearly straight down.

Spacecraft altitude       158.79
Emission angle              1.79
Sub spacecraft latitude    43.42
Sub spacecraft longitude  340.7


The width of the field is roughly 0.37 degrees of longitude, correcting that for the latitude, the width on the surface assuming small angles is:

$$w \approx r_{moon} \ cos(latitude) \ \frac{\pi \ 0°.37}{180°} \approx 8.150 \ (\text{km})$$

8150 meters divided by about 5000 pixels gives about 1.63 meters per pixel in y. And what does the data say...

Scaled pixel width   1.59
Scaled pixel height  1.55


So of course the scales in the metadata are going to be more reliable than my approximation, and you should use those numbers because a correct calculation would be much more complex! But my calculations serves as a) a sanity check (did I pass?) and b) an instructive illustration of what's actually going on.

Here is a quicker way to get at the scale, but it doesn't clue you in on the skewing. The table below shows that the LEFT camera has 10.00 microradians per pixel. Multiply that by the distance from LRO to the imaging location on the moon of 158.79 kilometers and you get...

1.59 meters/pixel! Same as the pixel scaling given.

For fun, you can go to https://lrostk.gsfc.nasa.gov/preview.cgi and see a real-time animation of what LRO is scanning right now!