[X] My Assistant

[Guest_DonPMitchell_*]

Here are images I generated from the 9-bit Venera-13 and Venera-14 data. Most of the work was spent combining three or four transmissions from the spacecraft, each with an independent set of digital noise. In some cases, scrambled regions of images were restored by recalculating the 10th parity bit, and shifting the bit stream. In particular, I resurrected a new section of the image on Venera-14 Camera II on the left side. I managed to distill out one very high quality copy of the full transmission from each of the four cameras.

Next, there is the problem of linearizing the camera response. The camera response curves published in Cosmic Research are wrong, or at least they do not extend into the darker range where a lot of the actual Venus imagery lies. You can prove they are wrong from the calibration wedges, viewed through the four different filters. Correct generation of true log response would result in wedge profiles that are exactly offset from one another. Some recent work on camera self-calibration in the computer-vision community points the way to reconstructing response curves, and when applied to the Venera images, the result is very pleasing. Round objects, like the elbow joint of the penetrometer, look round, not flat, details in shadows appears out of the blackness of the original Russian images, and some additional hills on the horizon appear out of the formerly white sky.

The full transmission consisted of several passes of the camera scanner, back and forth, across the scene. These four panoramas are combinations of up to five black-and-white images (clear filter), and a number of red, green, and blue-filter images. In Lab color coordinates, I extracted the ab channels from the red/green/blue images, and added them to the much higher quality B/W images. You can see that when making scans through the clear filter, the camera covered a wider area, the uncolored regions are just where the RGB data did not exist. Most of the blue images are black, due to a sudden drop-off in the camera response. There are probably a few areas near the bright horizon where the real RGB ratio can be extracted...a project for someone someday.

I've been too busy with my book and my company in Seattle to completely finish what I wanted to do. The color is still not correct on any Venera surface images. But the color filters in the camera were balanced with gray filters to be somewhat correct. I am awaiting one last key piece of data -- the spectral response of a color filter that was in front of the calibration wedge. With that in hand, an absolute color calibration would be possible.

Venear-13, Camera I (short program):



Venera-13, Camera II (long program):


Venera-14, Camera I:


Venera-14, Camera II: