I wondered this too
The Mars rover has a really shitty camera. If we’re going to Mars, I want color.
Remember, I do this to entertain me, not you.
Uncle Pays the Bills
 |
|
Find Local
|
Categories
- ATF (769)
- Death Match (10)
- Eminent Domain (225)
- In the bag (9)
- Notes to Junior (546)
- Poker (74)
- Recipes (142)
- Shotgun Noob (11)
- The Airing of Grievances (19)
- uncategorized (47,725)
August 9th, 2012 at 8:55 pm
When I left the military in 1992, we were still using vacuum tubes in some of our amplifiers. The government procurement process is awful.
August 9th, 2012 at 9:02 pm
I’m a digital camera design engineer working primarily with military, scientific, and industrial cameras. The camera on the rover actually uses a very good CCD. It’s a Kodak KAI-2020, which is indeed a 2MP camera, but resolution is a very over-rated aspect of camera performance. Of far more importance in these kinds of applications are aspects such as the efficiency at which the sensor converts light to electrons (quantum efficiency), low readout noise, good full well capacity (determines the maximum light level that can be captured), etc. Sensor pixel count is not only not that important, but usually isn’t even the determining factor in controlling how detailed an image the camera as a whole can capture anyway – the lens is far more important. Going to a higher pixel count sensor would have required higher power, more weight and size, all of which are extremely important on a space mission, and likely wouldn’t have made much of a difference in the image quality if any.
Some other things to consider – the pixel size on that sensor is pretty high compared to any consumer camera, so even though it doesn’t have so many pixels, each pixel is far better at capturing light. The larger pixel size also makes the sensor less sensitive to radiation damage, which is very important in a space going application. At one point a consumer sensor company in Japan had to stop using air freight to ship their sensors to the US because the sensors were developing too many bad pixels as a result of their exposure to the increased radiation at high altitude. They had to use a ship to bring them to the US. Obviously, that would cause issues on a Mars rover…
There’s lots of other reasons why that sensor is a good choice. Suffice it to say, though, that trying to apply a consumer electronics perspective to the design of scientific cameras is a really bad idea. The saying “they don’t know what they don’t know” comes to mind.
I’ve built several cameras around that same family of chips. One is a 1 MP camera that’s used in digital medical x-ray systems. Another was a 16 MP camera used for scanning film and aerial photography (that sensor is physically huge), and another is involved in a military platform that gets a lot of press. That family of sensors (now split off from Kodak as Truesense) does extremely well. Any of the commercial high megapixel count devices that most people are familiar with would fail miserably in similar situations.
August 9th, 2012 at 9:02 pm
generally in the world of engineering you get some choices, one of them is reliability vs performance. The more cutting edge you make the tech, the less you can count on it and the harder it gets used the faster it breaks. Going to mars, landing and then the environment there is pretty harsh, and a crappy camera that you are 100% sure will work is better than a fancy camera that only has 80% chance of working. How bad would it be to have a fancy working rover on mars but then the camera is misaligned. The geek squad doesn’t go to mars (yet).
August 9th, 2012 at 9:29 pm
Who cares how much resolution the camera has. The JPL team put a gas chromatograph and an X-Ray diffraction detector on a rover on MARS! Oh, and each instrument can operate at 100-120 Watts. Not too shabby.
August 9th, 2012 at 9:42 pm
They could have strapped a little Canon G-10 on it for $300. But who needs those huge, poster-sized, 15MP images anyway? If the images were satisfactory, there’d be no need to argue for another mission with a bigger budget next time. Nobody wants to go on umemployment, do they?
August 9th, 2012 at 10:03 pm
Sorry, this isn’t one of those “damned government!” moments.
First, we’re pretty familiar with Martian terrain now. It’s kind of rusty red with a shit load of rocks and not much else.
Two, there’s limited data that can be transferred per day, and there’s MUCH more interesting data than pictures as far as science goes. This thing is a laboratory on wheels, not a video rig for the next Michael Bay movie.
Three, as mentioned earlier, MP is a very, very, very misleading concept. I had a 4MP camera that blew away most 16Mp consumer grade crap. The sensor was amazing and captured light better. Blow up your pictures from your phone camera and you’ll see it’s interpolated (it ‘guesses’ what color would have been in the middle of the pixel).
Four – take your iPhone. Launch it on a rocket into space. Land it on the surface of Mars. A) you’ll void your warranty and B) it won’t survive. The reason our cameras are so cheap is because they don’t need to be hardened against radiation or physically strong enough to survive launching & reentry. Because they’re cheap, they’re disposable and companies can rush to make the next bigger thing without losing anything. The camera on this thing needed to be a bit better suited for the job and well tested.
Five – In regards to space, I’m a complete and utter hypocrite. Allow me to get out of my wookie suit because as wasteful and bloated as our government is, I still squee like a little girl with every picture and report from these things. Given the power to change things, I’d not divert the money to things like this, but I can’t help but sit in awe as I realize my children have never known life without pictures streaming from the surface of Mars. And in 2 more years or so, we’ll actually be able to see close up shots of Pluto. Society will collapse around our ears, but we’ll have great pictures of the planets!
August 9th, 2012 at 10:11 pm
What, nobody is going to mention the huge increase in bandwidth required to send back color vs b&w pictures? More electricity to take the pictures, more time to transmit (which would take even more electricity), etc.
August 9th, 2012 at 11:44 pm
The rover has been running on a program (think of it as an operating system) since launch last November that was optimized to run the landing process and take some quickie grab shots, just in case things went very badly.
Starting Saturday, JPL is uploading a new OS designed for surface operations, and they’re putting science on hold for the four (Mars) days it’ll take to get the upload done and tested.
THEN you’ll see a lot more imagery, both grayscale and color.
August 10th, 2012 at 12:03 am
People make it sound like you just need to duct tape an iPhone to the rover get better pictures. Between the vacuum and radiation in interplanetary space, it’d be dead.
An IBM PowerPC 601 processor costs $500
An IBM RAD6000 (the radiation hardened version of the 601) costs $300,000
Commercial satellite manufacturers wouldn’t be paying a 60,000% markup if the consumer version would still work in space, so comparing the performance of what’s possible in space-ready technology to what’s possible in consumer technology is a pointless metric.
August 10th, 2012 at 7:39 am
There are seventeen separate cameras on Curiosity, some of which are paired into stereo pairs.
http://marsmobile.jpl.nasa.gov/msl/mission/rover/eyesandother/
August 10th, 2012 at 8:29 am
OK, MikeR beat me to most everything I was going to say. But I will add that you cannot fix something on Mars. It either works, or it does not.
Same goes for any remote sensing platform. Here on Earth we are lucky enough to frequently get them back, so we can eventually fix them. But even in those cases, we’re looking at some kind of mission fail. Unfortunately, when you need it, you REALLY need it. That means we build conservatively and carefully. Measure twice, cut once.
I’ve worked as a “firefighter” in the defense/space industry. That means by the time we show up, projects are over-budget, under performing and late. Some of these programs have a needed capability, so they call people like us to try to recover something of value. People like us are the last ones to show up before the project gets cancelled (and sometimes we show up to shut it down…that’s never fun).
The common denominator in most of these failing programs is the ever-present desire to go too advanced and use the most modern “stuff”, whether it be software, hardware or “practices”. Then comes scope-creep – the temptation to do a little more than you must, because it would be “easy” or “low-hanging fruit”. You have no idea how many times I have heard the word “cool” used to describe a failing idea. Rule One: eliminate “cool”…
There is truth to the all-too-slow engineering process used today in US systems. I won’t defend it. I hate it. The entire “Systems Engineering Approach” is broken beyond repair. At the end of the day, every single successful program (including ones that were rescued) are eventually successful because of a tiny group of dedicated and smart people who just make things happen. They frequently labor in obscurity until management figures out that their “practices” are the problem…then they let the good guys loose to do their thing.
Anyway, let me end this by saying I think that all those who pointed out that sometimes ‘less is more’ are correct. The trade-off between a functional 2004 sensor and non-functioning 2012 sensor is not a trade-off on planet Mars. There is a whole other dynamic in the mind of someone building these things…we won’t trust it until it has broken (and been fixed) many, many times.
FWIW, I have personally designed custom sensor systems – big and small – for many years. It’s good work, but rare to find new people willing to do it anymore. If you have an interest in it, please consider it. Our industry needs the help and frankly it’ll keep you busy and well employed for a long, long time.
August 10th, 2012 at 9:44 am
Let me add a +1 to everything Patrick said. Space is harsh on our equipment, especially when you get far enough out that earth no longer offers any protection from radiation.
I’ve only worked with imaging Earth from space, but even that is more difficult than it seems. And once something is launched… it can’t easily be fixed. Other than the ISS and Hubble I can’t think of anything NASA has sent people to repair.
August 10th, 2012 at 11:33 am
HD color pics are supposed to be released this weekend or early next week. Each picture takes HOURS to download due to the amount of data.
August 10th, 2012 at 11:40 am
They could have strapped a little Canon G-10 on it for $300.
Consumer electronics are promptly going to fry in the Mars radiation environment. We have it very good here on earth, our magnetic field and atmosphere protect us from an awful lot of shit the sun spits out. Radiation levels are much higher on Mars due to thin air and a lack of magnetic field. You can also get wild temperature fluctuations on the red planet, and the electronics have to hand that too.
August 10th, 2012 at 11:40 am
Commercial satellite manufacturers wouldn’t be paying a 60,000% markup if the consumer version would still work in space, so comparing the performance of what’s possible in space-ready technology to what’s possible in consumer technology is a pointless metric.
This!
August 10th, 2012 at 12:58 pm
“And in 2 more years or so, we’ll actually be able to see close up shots of Pluto.”
So long as we never have to see close-ups of Neptune.
August 10th, 2012 at 1:39 pm
As a practicing engineer, I can’t add too much to what’s already been said other than when you can’t send out a repair team, reliability is the #1 concern. That means you set a specification and you stick with it for as long as it takes in order to have sufficient integration and test time.
August 10th, 2012 at 2:21 pm
I think it was Newt G who posed “how much farther along in space exploration would we be if NASA were willing to take the same level of risk as NASCAR?”
August 10th, 2012 at 2:24 pm
I’ve already seen a color panorama pic from the rover. pretty badass.
August 10th, 2012 at 3:28 pm
“So long as we never have to see close-ups of Neptune.”
And especially no close-ups of Uranus.
*rimshot*
August 10th, 2012 at 3:32 pm
Robb Allen:
I disagree. This is a damned government moment.
This ain’t the 1960s.
Buck Rodgers is what brings in the bucks. If they cannot produce photogenic images for the people who are actually paying for this, they won’t get money to fund the next one.
The scientists might not like it, but that is just too damned bad.
August 10th, 2012 at 3:41 pm
Sometimes you just have to hire a nerd-wrangler, I guess.
They have no fucking clue about how to deal with their real customers, the Public.
August 10th, 2012 at 3:42 pm
Does this damned camera have a color filter wheel?
If it does, it can be used to take at least one good photo each day to sell the program with.
August 10th, 2012 at 3:44 pm
It would have been nice if they could have put a camera on there that they could tell from earth to take a photo in black and white or color.At least then you could send a couple of nice full color images down.
August 10th, 2012 at 4:35 pm
HL, wrong week to compare NASA and NASCAR. I remember NASA before Chrysler pulled out of that, too…
You can’t send a repair crew? The hell you can’t.
My son is a satellite designer, and I told him if he shows up with a Mohawk with little stars on the side, I’m not leaving him the good slide rule like I promised.