This camera system is better than lidar for depth perception

Those techniques are far from perfect, and can produce significant errors. Locating homologous points is, in general, non trivial.

Given that I'm significantly deaf, anything that might blind me is of concern, so I would prefer not to have lasers in the environment.

Camera systems won't do much good in fog or other bad climate situations, so I guess Lidar is here to stay

On a sunny day (Sat, 30 Oct 2021 00:38:58 +0200) it happened Klaus Kragelund snipped-for-privacy@hotmail.com wrote in snipped-for-privacy@nntp.aioe.org>:

Do you think you should be driving that car in fog so thick you cannot see anything? And you can perhaps go to longer wavelength:

Lidar gives you positive information of ranging. Cameras infer it from images that are subject to constantly changing lighting conditions such as sun flare, dusk, nighttime, rain and snow (forget about dense fog - Lidar won't work then, either). I have driven on wet streets in nighttime where it was extremely difficult to see the lane stripes. If I can't see them then neither can a camera.

surely that depends on the camera

Really??? So you are happy letting your car drive through blinding snow or dense fog that you can't see through? That sounds like a 100 car pileup on the highway to me!

I don't normally respond to your nonsense posts, but this one is so problematic and easy to explain. Cameras can see what you don't. They are not subject to the same limitations and have much better control over brightness and contrast of the image being analyzed. An example is photography of people where the camera essentially sees through the weave of the clothing.

There are times when my auto pilot doesn't like the weather conditions, but it is not universal and is typically the same point or worse conditions than when I am able to drive the speed limit.

Don't assume a camera "sees" the same things you do.

On a sunny day (Sat, 30 Oct 2021 10:44:47 -0700 (PDT)) it happened Flyguy snipped-for-privacy@yahoo.com wrote in snipped-for-privacy@googlegroups.com:

You can have a camera with a much wider spectral range than the human eye, starting with infrared for most of those (those normally have an IR blocking filter). I have some Sony starlight 0.01 lux ones, you'd be amazed what that makes out at night, mine actually is also IR sensitve.

lørdag den 30. oktober 2021 kl. 22.32.11 UTC+2 skrev Jan Panteltje:

IR is useful for seeing through mist, but not fog or rain. Neglecting absorption, the basic issue is the size and density of the particles compared with the wavelength of the EM radiation.

Droplets large compared with a wavelength (e.g. heavy fog or rain) scatter all wavelengths the same way. Very small particles have a scattering cross-section that goes like (volume/lambda)**3, which gets small pretty fast with increasing lambda.

Cheers

Phil Hobbs

But down at that level you have to integrate for awhile to get an actual image, whereas a driver has to respond fast enough to avoid hitting things.

Cheers

Phil Hobbs

None of that matters. When the rain is so heavy visibility is not the only reason to not be driving. There are issues with rain not flowing off the road surface fast enough creating hydro-planing hazards. Bottom line is if the visibility isn't there, you probably should not be driving anyway.

My Tesla could not continue with autopilot once because the snow covered the cameras near the bumper. That was because I was cresting a hill with the wind blowing wet snow which made visibility very poor and stuck to the cameras. Autopilot disconnected and the road got slippery so I had to drop speed to 30 mph and less.

Cars have headlights and don't need to do such integration. So this argument is specious.

On a sunny day (Sat, 30 Oct 2021 18:10:46 -0400) it happened Phil Hobbs snipped-for-privacy@electrooptical.net wrote in snipped-for-privacy@electrooptical.net:

Sure, but these cameras are still pretty fast in fps, not the usual CMOS sensors but CCDs almost usable for astronomy. And to detect 'life forms' in IR crossing a road these cameras are better than the human eye, the body heat shows up as bright white, There are youtube videos with one from a drone flying at night. And of course you can add some IR beam if needed, no need for a focussed scanning laser, many security cams have a set of IR diodes around the sensor.

many videos with this sort of camera for FPV.

Some specs of the ones I have (pdf):

mine have a different mount with and I have all sort of lenses including zoom and wide.

And last but not least for targeting heat from airplanes

Price of a drone versus price of an F35 Good Buythen

On a sunny day (Sat, 30 Oct 2021 14:28:45 -0700 (PDT)) it happened Lasse Langwadt Christensen snipped-for-privacy@fonz.dk wrote in snipped-for-privacy@googlegroups.com:

Exactly, nice example!

Personally I have always been against LIDAR, do not like lasers shining around.

IR is useful for seeing through mist, but not fog or rain. Neglecting

Fog is common here.

Sometimes it is widespread across many miles.

Sometimes it is local; I frequently see that out of my kitchen window; it is always very pretty.

Sometimes it it extremely small area, often where a road crosses a river in a valley.

Hence it is not uncommon to _suddenly_ encounter fog, particularly at night. Even when drivers are paying attention (unlike in a Tesla etc), they are caught out and crashes happen.

Personally I have always been against LIDAR, do not like lasers shining around.

IR is useful for seeing through mist, but not fog or rain. Neglecting

How is your car driving /today/? Will it recognise fo or not?

Lidar has the same problem in fog. It has a small advantage in mist, where the longer wavelength helps because the droplets are very small, but in fog, rain, snow, etc., the scattering is geometric and hence wavelength independent (apart from absorption).

Cheers

Phil Hobbs

You are not caught off guard by this in a Tesla. It sounds an alarm and lets you know with no uncertainty!!!