Something capable of putting enough light to make a "this is a cycle lane" image that's visible in daylight is going to have to be much brighter than a laser pointer.

The illumination level on an overcast day is about 1000 lux, so to be visible in daylight the device will have to light up the bright bits of its image at least about that brightly.

The image in the picture looks to me like the illuminated area is at least about 0.1m^2, probably rather more. So that's a total output of at least about 100 lumens.

Light at the peak of the photopic curve (i.e., wavelength optimized to look as bright as possible per unit power) is 683 lumens per watt.

So the output from this thing is going to have to be at least about 150mW. That would be well into class 3B, meaning that you need a key switch and safety interlock on the device and protective goggles if you're anywhere where it could shine into your eye.

(Doing laser safety calculations properly is complicated. Don't take the above too seriously.)

When you assumed the illuminated area to be 0.1m^2, you assumed the light source to not be a laser, spread evenly over a 30×30cm square (to get 1000 lux). Realistically, a several-watt fluorescent bulb would throw 100 lumens. And would of course not be harmful to look at.

The point of a laser is that the light is /not/ scattered over a square, its a very tight circle. If it was instead over a 3mm diameter circle, that 100 lumens would be over 20 million lux. You would, of course, not want to stare into that!

Its unclear from the picture exactly how long the "line" being drawn is (3mm diameter by even 7m long is only 0.02m^2), but a hypothetical kid staring into the beam would only get hit by part of the figure; the laser beam is moving to project the entire figure. Only the part that actually goes into the kid's eye matters, and the impact is lessened by the fact that its moving and also by the blink response.

BTW: Wikipedia tells me direct sunlight is up to 130,000 lux, so that's probably around the illuminance you need, not 1,000 lux, btw.

I didn't assume that anything is spread evenly over a 30x30cm square, nor did I assume that the light source isn't a laser. I assumed that the optics were arranged to get the light from the laser distributed in the appropriate way. (There are lots of ways to do this. You could do it with a single DOE, which would probably be simplest if you didn't mind poor image quality; or with a few lenses to give you a nicely diverging beam, and then something to absorb light where you don't want it -- but that would waste a lot of power.) It doesn't stop being laser light, and it doesn't stop being subject to laser safety regulations, just because it isn't coming out as a nice collimated beam.

The point of a laser is not only that the beam is very small. It's also that it's possible for the light to be focused on a very small region of the retina, even if the beam is diverging when it enters the eye.

The illuminance needed for the figure to be clearly visible is of course an average over time (roughly, the integrating time of the eye). If you use a scanning system (which seems a bit of a funny choice for something projecting from a moving object), then indeed the beam may only be in your eye for a small fraction of the time -- but it will also be much brighter (while it is in your eye) than I assumed. The way in which laser safety changes when you consider a pulsed system rather than a continuous one (which is what you effectively get in this scenario) is complicated, but generally for a given average power it's worse to have it delivered in pulses. In any case, though, the critical scenario for a laser safety assessment is when someone manages to get close enough to the laser to get rather a lot of the beam into their eye. (You can have safety mechanisms that aim to make that impossible, of course.)

For laser classes higher than class 1, the blink response is already taken into account in the permitted exposure levels. (Class 1 safety, which this device would be miles away from having any chance of achieving, is extremely strict and basically amounts to "if you are really determined to damage your eyes using this device, the best way would be to hit yourself in the face with it". To be class-1 safe, you need to be safe even without the blink response. But that's not the issue here.)

I chose the figure of 1000 lux precisely because I wanted a low estimate, to make the point that even being very generous to this thing it would be difficult to do it without a laser powerful enough to raise safety issues. Obviously if you want the image to be visible even in really bright direct sunlight then the numbers need to be considerably higher.

I may be missing something, but when you take 1000 lux and 0.1m², and come up with 100 lumen, the obvious way to put those numbers together is 1000lux × 0.1m² = 100 lumen. That's where I got evenly spread over a 30cm square (30 being the square root of 100, to one significant digit).

I'm not an expert in the field by any stretch of the imagination, but something seems very wrong with these calculations in that throwing a several-thousand lux image of that size would be easily done with a $500 business projector.

Yes, I assumed that the light is spread over that area. I didn't assume it's spread evenly over a square of that area. Maybe the shape of the 0.1m² region wasn't important to whatever you had in mind, but then I'm really unsure what your point was since I'd already said I was assuming an area of 0.1m².

You can get very bright $500 business projectors. They aren't emitting laser light. If they were, they would have serious problems getting non-scary laser safety classifications. Part of that is because the rules governing laser safety classification are kinda paranoid. Part of it is because coherent light really is specially dangerous. (A 100W light bulb has a luminous efficiency of about 2.5%. A 2.5W laser is really, really, really dangerous.)

Anyway, you could do your eyes some serious damage with that $500 business projector. It just happens that there aren't such stringent regulations governing incoherent light sources, because the worst case isn't quite so scary as it is with lasers.

It's very hard to get blinded by a laser, but it's quite easy to get dazzled for a few seconds if one goes in your eye. Being dazzled for a few seconds usually isn't that dangerous unless you're, say, driving a car.

No sir, I don't like the idea that the inconsiderate laws-don't-apply-to-me cyclists I carefully avoid running over every day will suddenly be equipped with lasers which can shine into my eyes by accident.

But to get dazzled the viewer would have to be between the from of the bike and the ground in front of it. I don't imagine many car drivers would be there, but it a concern with children passing in front of the bike when it isn't in motion, say.

Being between the front of the bike and the ground isn't hard to imagine if the cyclist is cresting a hill.

Assuming the device is 2ft from the ground, and projecting an image 10ft out, that'd be mounted at an angle of 79° (with 90° meaning parallel to the ground). If you're coming up a symmetric 5° hill, and the beam goes over the crest, it'll go 115 feet before hitting ground. 6° hill, and it'll not hit ground until the descent ends.

(I have no idea where the device is mounted on the bike, or how far in front you'd have to project. Easily could be further up than 2ft, which would help, or further out than 10ft, which would hurt).

Or if the bike falls over.

It's a valid concern, but is it any different to dealing with a car with poorly adjusted headlights?

Very, very, very, very, very different. The lasers we are talking about are beams, which means they are not subject to significant intensity falloff over distance. Thus, the energy they put out at the source can be efficiently transmitted over great distances directly to your eyeball. This causes further problems when the laser can deliver enough energy to damage your eye, or even permanently destroy it, in less than the ~150 milliseconds it takes you to reflexively blink it. Run this thing over a bunch of broken glass and you will have legitimate problems with localized areas of dangerously high light intensity.

Also, lasers are not toys. Do not let children play with them. Do not let anyone "play" with the green ones or any stronger lasers. You are taking insane risks when you do; yes, you can probably keep it out of your eyes but it only takes a moment's inattention and you can be talking serious lifelong consequences. It's a poor tradeoff for entertainment purposes.

Not only is this design far, far too dangerous to let out into the public in a big way, even the testing was far more dangerous and hostile to the people in the area than the engineer realized.

I've only been hit in the eye with a laser a couple of times, but I'd say yes.

And I've zapped myself several times deliberately, even holding the weak little class IIIa laser pointer right up to my eye to show people that the "dangers" are too often overstated. Granted, that's only for low-wattage laser pointers. There are high power lasers where the "do not look into laser with remaining eye" warning is absolutely correct, after all, as well as a lot of different things in between. Lasers aren't inherently dangerous after all simply because they're lasers, it's a matter of how much energy they're zapping your eyes with and for how long.

So I'd have to say, no. If reasonably engineered, this would be less likely to be harmful than someone's headlights shining in your eyes. They're so focused that they wouldn't hit you in the eye for very long given that you're a moving target likely to be several meters away from the bike.