Alien Anomalies

OBrien, the images you have posted are of shocking quality even to my untrained and uneducated eye. Was this intentional, as I am sure there are much better quality images that could prove your point ?

Also, while we are on the subject and considereing you are an expert on the subject I have two questions...

In your qualified opinion, is the quality of the images you posted (considering the angle of the light and the distance from the object) in keeping with the detailed design of the equipment you have so kindly explained to us ?

Simplified, is it your learned opiniion that the best engineering possible that was used for these cameras is delivering results that match the specifications of the design ?

Chandre wrote:

---

OBrien, the images you have posted are of shocking quality even to my untrained and uneducated eye. Was this intentional, as I am sure there are much better quality images that could prove your point ?

Also, while we are on the subject and considereing you are an expert on the subject I have two questions...

In your qualified opinion, is the quality of the images you posted (considering the angle of the light and the distance from the object) in keeping with the detailed design of the equipment you have so kindly explained to us ?

Simplified, is it your learned opiniion that the best engineering possible that was used for these cameras is delivering results that match the specifications of the design ?

---

Yes, I intentionally posted these images rather than others because these are the most accurate images available. I know that sounds surprising and against common intuition so here's an explanation.

The Pancam CCDs are 1024 x1024 pixels. When a picture gets taken the light falling onto a single pixel is reported as having a single intensity value. Pixel (1,1) has a single value assigned to it. So does pixel (437,836). All the way to pixel (1024,1024). That's all the image is that gets transmitted, a list of numbers associating a particular intensity value to a particular pixel.

Those numbers are reassembled into a grayscale picture. As I noted before, the grayscale is simply a convenient convention and a way we're used to seeing pictures. The intensity values from minimum to maximum could have been assigned any colormap --- black to white produces the image we see. If the colormap was white to black the image produced would have been a negative. If the colormap was rainbow it would produce a false color where dim was red and bright was violet.

Even in the case of grayscale, different colormaps are used. If the shade of gray is directly proportional to the intensity (linear) then you'll get a particular brightness and contrast. In many cases, a logarithmic mapping is used, because of the way the human eye responds --- something that's twice as bright in number of photons doesn't look twice as bright to your eye.

But the important thing is that there is no more information in the picture. That's all there is. If there are two tiny bright points very close together and the light all falls into one pixel, there is no way to distinguish that it's two points instead of one point.

When you look at a zoomed in section that will make the image look blocky because you're seeing the individual pixels that make up the 1024 x 1024 image. And no, it's not possible to press the "Enhance" button like they do on television shows to instantly clean up the picture. There are some techniques that can be used, but they rely on knowing the exact form of the object you're looking at. That is, the extra information comes from outside. For instance, if you know you're looking at a license plate that you know has numbers and letters, you can distinguish between an S and an 8, but you can't determine if a random blob is a squiggle or an 8.

Even though you think the image quality is shocking, it is the most accurate. It shows the difference in intensity between adjacent pixels.

Take a look at these two images, and decide which you think is "better".

Image 1


Image 2


Most people would say that Image 2 is better. But there's a difference between fidelity and smoothness. Image 2 is simply image 1 that I have blurred slightly. Image 2 has LESS information than Image 1 because I have mingled information from adjoining pixels and lessened the accuracy about the information of relative intensity of the pixels. Image 2 "looks better" but is not as accurate as Image 1. For instance, the parallel ridges in Image 2 all look very rounded and weathered. That's not necessarily the case, because I smoothed the image with blurring. They could be more sharp-edged, but there isn't enough information because the ridges are only 2-3 pixels wide. There's no way to know with this image.

And for figuring out the distances of objects using parallax with the left and right Pancam you need to know the differences in pixel numbers from the Pancam CCDs and not the pixels you get out of some random JPG. That's why Image 1 is useful for that purpose.

With the color image that's been used before, there is more information, because that's made up of 3 or more different images taken through different filters and all combined into one image. But filter set that gives the red-green-blue only exists on the right Pancam, so there's not an equivalent picture I could show that would be the same color composition from the left that could be compared to the one on the right.

Let me say this again in summary. While it seems contrary to common sense, Image 1 is MORE ACCURATE than Image 2, even though Image 2 looks smoother and more natural. Smoothness does not mean accuracy.

To answer your other questions, yes. This is exactly what I would expect from a 1024 x 1024 pixel CCD in a camera with this focal length. The limitations of the image in this case is based not on the lenses, but on the number of pixels in the CCD. It's essentially a 1 megapixel CCD for each camera, and it's taking a photo of a rock 35 feet away. If the camera had denser pixels, then the image would have better resolution.

Veteran Member Posts: 669 Date: Mon Aug 30 00:30:10 2010

Reply Quote More  Edit Post Delete Post Printer Friendly Ban IP Report Spam That explanation is clear and understandable. However, it does not address the compression artifacts and manipulation that are in the photos we see. The ORIGINAL may have all the information, but the one we get to see has blacked out areas and smudging and the jpeg compression artifacts. All this makes what you say TECHNICALLY correct but PRACTICALLY does not help us to arrive at the truth and to see what was in the original picture. But the important thing is that there is no more information in the picture. That's all there is.If you are talking about the whole of the image from which this section comes, this is not true. It is not true because there IS more information in this picture but it is not available to us because of the blacked out areas, the smudging and the compression artifacts. It is misleading to make statements like this when it is obvious that, due to circumstances beyond our control we cannot arrive at the original information in the picture - unless you can point us to a way of arriving at that information. Can you? __________________ Search for Mars Rover Photographs at http://mars.pelicanbill.com

Chandre Veteran Member Status: Online Posts: 835 Date: Mon Aug 30 14:00:15 2010

Reply Quote More  Edit Post Delete Post Printer Friendly Ban IP Report Spam Thank OBrien for that technical explanation and I think I have the gist of it. In essence you are stating : This is exactly what I would expect from a 1024 x 1024 pixel CCD in a camera with this focal length. The limitations of the image in this case is based not on the lenses, but on the number of pixels in the CCD. It's essentially a 1 megapixel CCD for each camera, and it's taking a photo of a rock 35 feet away. If the camera had denser pixels, then the image would have better resolution. Perhaps NASA should work more closely with Sony Ericsson in future missions. Their little mobile phone camera is able to deliver the following (and in colour)......... Camera Operation and Options: The view button turns the camera to review mode while the shutter release button returns it to photo mode. The camera has four shooting modes: Normal, BestPic, Panorama, and Frames. Scene mode options cover almost every event: Auto, Twilight Landscape, Twilight Portrait, Landscape, Portrait, Beach / Snow, Sports and Text. In playback photos can be shown individually or as 9 or 25 thumbnails. Screen / LCD display in photo mode: The screen has a resolution of 320 x 240 pixels and updates fairly smoothly - the colours appear accurate for reviewing photos. There is no live or review histogram available but the screen is clear and the text and menus are easy to read. Optical Viewfinder: There is no optical viewfinder. Shooting Options: Mode, Scenes, Picture Size, Focus (Auto, Macro, Infinite), Flash, Self-Timer, Effects (shown above), White Balance (Auto, Daylight, Cloudy, Fluorescent, Incandescent), Metering mode (Normal, Spot), Picture Quality, Shutter sound (choice of four), Turn on time and date, Reset file number, Save to (Memory stick or phone memory). Photo shortcuts screen: By pressing the 0 button on the numeric keypad whilst in photo mode, the camera will display the photo shortcuts screen shown above. This provides quick access to photo scenes (top left icon / button), photo modes (bottom left icon / button), zoom, picture resolution, focusing mode, night stabilisation, and flash. Playback mode: Scrolling through the photos is very quick. The zoom is fairly quick and works at up to 32x. Very basic shot information is shown about the images. From playback mode you can Print, Edit in PhotoDJ, Rotate, Delete, View a Slideshow, or Send the picture (as an MMS message, or as an email). Picture Size / Quality: The camera takes the following size pictures, and the following number of pictures will fit in the provided memory: Image Size: Number of Photos Stored / Quality Fine Normal 3M (2048x1536) 78 100 2M (1632x1224) 124 217 1M (1280x960) 242 300+ VGA (640x480) 300+ 300+ As shown in the table above, higher quality images take a larger amount of memory, however, as the camera is fairly low resolution compared to other digital cameras, you can fit a lot of photos in the provided memory, assuming you don't use the memory for other files (mp3s, videos etc). If you are only interested in small photos, then you can fit a very good number of 2mp, Normal photos in the memory. The K800i has 66mb of memory built in and assuming you don't fill the shared memory with videos and mp3s, then you should be able to fit upto 100 3MP, Normal photos. There is the choice of four image sizes: 3mp, 2mp, 1mp and VGA, and there is a choice of two compression options: Normal or Fine. Unfortunately there is no choice regarding aspect ratio. You can choose between "High Quality" video or low quality video, the later is designed for sending as MMS text messages or attached to emails etc. If you need to fit more mp3s, photos or videos then you can expand the memory with a Sony M2 memory card, prices are around £20 for 512mb or around £40 for 1gb assuming you shop around (and assuming you can find a shop that sells them. Source : http://www.digicamreview.co.uk/sony_ericsson_k800i_review.htm ......it may save NASA a lot of money and we could have colour images on a camera that could possibly even 'phone home'. LOL

OBrien Veteran Member Posts: 112 Date: Mon Aug 30 16:39:58 2010

Reply Quote More  Edit Post Delete Post Printer Friendly Ban IP Report Spam Chandre wrote: Perhaps NASA should work more closely with Sony Ericsson in future missions. Their little mobile phone camera is able to deliver the following (and in colour)......... Perhaps you could point out to me the last time a Sony Ericsson mobile phone camera survived a 500 million mile interplanetary trip and then worked for five years in the conditions on another planet. I'd buy that phone! And if you're still unclear about why the Pancam CCDs don't have little color filters in front of every pixel, like a cell phone camera, I wrote a little tutorial that you might find helpful.

Chandre Veteran Member Status: Online Posts: 834 Date: Mon Aug 30 18:20:16 2010

Reply Quote More  Edit Post Delete Post Printer Friendly Ban IP Report Spam Perhaps NASA should add the phone to the arm extension and send it to see if it would. Mines survived accidental submersion in my bath and being dropped down a flight of stairs.... considering that they are built to survive being constantly fiddled with by heavy-handed teens I have no doubt a little trip to Mars would be a walk in the park. I read the tutorial, but to be honest I still don't understand the logic. If Sony can fit a phone and a superior camera into a small (in comparison) machine why can it not be done in a rover with the best minds in the world and all the money at NASAs disposal to make it work more effectively than their current option ?

Veteran Member Posts: 111 Date: Mon Aug 30 19:48:22 2010

Reply Quote More  Edit Post Delete Post Printer Friendly Ban IP Report Spam "'I have no doubt' it would work" is generally not considered sufficient validation testing for spacecraft hardware. There's usually a lot more analysis and testing involved. That's why experienced engineers who understand the specifications, constraints, and environments of the system are given the task. That way it has a much better chance of working.

OBrien Veteran Member Posts: 110 Date: Tue Aug 31 03:49:01 2010

Reply Quote More  Edit Post Delete Post Printer Friendly Ban IP Report Spam thewatcher wrote: I suggest OB opens a thread on image quality.. as i dont really want that subject to drown out the hunt around Payson Crater. Sorry for the unintentional straying from the topic. I was responding to direct questions posed to me.

qmantoo Veteran Member Posts: 668 Date: Tue Aug 31 04:20:25 2010

Reply Quote More  Edit Post Delete Post Printer Friendly Ban IP Report Spam O'Brien - and while you are at it, perhaps you could respond to my points in my post above too please? That would help us get to the bottom of this matter. 1) ...it does not address the compression artifacts and manipulation...The ORIGINAL may have all the information, but the one we get to see has blacked out areas and smudging and the jpeg compression artifacts. 2 ...there IS more information in this picture but it is not available to us because of the blacked out areas, the smudging and the compression artifacts. It is misleading to make statements like this when it is obvious that, due to circumstances beyond our control we cannot arrive at the original information in the picture - unless you can point us to a way of arriving at that information. Can you? __________________ Search for Mars Rover Photographs at http://mars.pelicanbill.com