Experimental headlight system can see through rain and snow

Experimental headlight system ...
A lab test illustrating how the new system can make raindrops "disappear"
A lab test illustrating how the new system can make raindrops "disappear"
View 2 Images
The lab setup, consisting of the projector, camera and beamsplitter
The lab setup, consisting of the projector, camera and beamsplitter
A lab test illustrating how the new system can make raindrops "disappear"
A lab test illustrating how the new system can make raindrops "disappear"

Driving at night in falling rain or snow can be treacherous, but not just because the asphalt is slippery – visibility is also greatly reduced, as the driver’s view of the road ahead is obscured by brightly headlight-lit raindrops or snowflakes. In the future, however, that may not be so much of a problem. A team led by Carnegie Mellon University’s Prof. Srinivasa Narasimhan has developed an experimental headlight system that renders most foreground precipitation virtually invisible, while still adequately illuminating the road beyond.

At the heart of the system are a digital light projector (which serves as the actual headlight) and an adjacent video camera. The camera is able to “see” the projector’s exact field of illumination, via a beamsplitter. As a raindrop falls into the top of this field, it is illuminated by the projector and its image is picked up by the camera.

A microprocessor then calculates the drop’s trajectory, and proceeds to selectively deactivate the projector’s light rays along that path. The result is that the raindrop is able to fall through the field of illumination, but with no light rays actually striking it – except for at the very top of the field, as it’s first detected. All of the other rays, or at least those that aren’t lined up with a raindrop falling within three to four meters (10 to 13 feet) in front of the projector, proceed through to light up the road.

The lab setup, consisting of the projector, camera and beamsplitter
The lab setup, consisting of the projector, camera and beamsplitter

This process is carried on for multitudes of drops simultaneously. For each individual raindrop, the amount of time between detection and reaction is approximately 13 milliseconds. Because the light rays are turned off and back on so quickly, there is reportedly no noticeable “flickery” quality to the light.

Needless to say, because numerous rays from the projector are continuously being disabled, it isn’t able to illuminate the road quite as brightly as would be possible otherwise ... although it’s not as pronounced of an effect as one might think. Even in heavy rain, the air volume consists of only about two to three percent raindrops, so the projector’s light output would only drop by about the same amount.

According to a report in Technology Review, the system is able to “hide” 70 percent of raindrops in simulated thunderstorm conditions, at a driving speed of 30 km/h (18.6 mph). That figure drops to 15 to 20 percent when the speed is increased to 100 km/h (62 mph).

Source: Carnegie Mellon University via Technology Review

Very impressive but so very complex, requiring tremendous amounts of calculations. The examples shown only have a single curtain of raindrops. What happens in reality when the rain has depth? And in heavy rain? Besides that, there is the added reality that electronics and rain (humidity) don't mix. To come up with a reliable system based on this principle may take a very long time.
I hope a simpler, more fundamentally correct solution comes about.
Nantha, I hope you don't drive your car in the rain... I wouldn't want you to break down when one of the engine management electronic sensors or control modules get wet.
Mel Tisdale
Phew! I imagine the cost of a simple bump into the tail end of another car in traffic would ensure that you forever drove with double the safe stopping distance from the vehicle in front.
In passing, Edwin Land, he of Polaroid fame, had the brillian idea of making all windscreen glass vertically polarised and all headlamps horizontally polarised (I might have them reversed, but it doesn't matter). The result would much safer night driving due to oncoming drivers not being able to dazzle the driver.
Seeing as the above process is controlling reflected light and reflected light is always polarised, if I remember right, then Mr Land's idea might also be effective in cutting reflected glare from rain drops etc. and be a cheaper solution to the overly costly and complicated system being proposed.
Except for very expensive top end cars I do not see this being embraced in cars. However, the aviation industry will be very interested. Even if it just allows for safer travel in the airport lanes. Industrial ship industry may be interested.... The large truck industry may also be interested...
Now this is a proper application of scientific principles - well done.
Bruce H. Anderson
As I understand it, light is polarized parallel to the reflective surface, such as the water in a lake or stream, or even aphalt. But since the drops are round, with no flat surface, would the reflected light be polarized? If it was, then some inexpensive glasses or a flip-down visor would do the trick. I don't know if the adverb "overly" applies here, but it is definitely costly and complicated, at least in this early stage. And I wonder what the effect water will have on the lens of the camera and its ability to "read" the location of the drops. Still, quite fascinating.
What would be a good idea, is to drive more slowly in rain and snow. I had the polarising idea (after Edward Land) It does work. The drawback is, the polarising filter darkens the window, so not so good after dark. Why not try looking at rain through some Polaroid sunglasses, at night?
Hmmmm....What happens when the wind randomly blows the rain around ??? I personally find that it is the rain on the glass that causes the most problems. There are good windscreen/windshield coatings out therethat you can buy that help immensly. What about re-inventing wipers or using ducted fast moving air to stop rain getting to the glass in the first place !!
Gary Richardson
Perhaps dual cameras will be incorporated into phones and tablets to offer a safety illumination app on camping trips? Futher upgrades to apps will offer improvements in illumination economically due to a larger customer base to divide costs. Additionally, increasing processor speeds bring the evolving benefit of error correction and predictive algorithm improvements.
Waterproofing the electronics is also no problem because cellphones can now be sealed from water already and if you visit this site frequently, you have already seen solutions to the problems commented above.
Gary Richardson
The more I read this the better I like it. More visibility means increased probability of averting dangers.