Computers giving cyclists a "right hook" sooner than expected

On Thursday, February 2, 2017 at 6:07:27 PM UTC-5, Frank Krygowski wrote:
On 2/2/2017 4:44 PM, Sir Ridesalot wrote:
On Thursday, February 2, 2017 at 3:11:58 PM UTC-5, Frank Krygowski wrote:
Snipped

I wonder how well those things work. They're said to use thermal
imaging, IOW infra red, I suppose. But ISTM that at certain ambient
tempreratures there might not be much temperature difference between the
bulk of a cyclist's image and the surrounding road. That would depend
on choice of outerwear. A jacket over a sweater, for example, might be
the same temperature as the air, especially in front where it's being
cooled by the relative wind. Helmets would certainly be at ambient
temperature on their outside. Cyclists with rain capes or with front
loads might have their legs hidden from such a camera.


There is a HUGE difference between Therma Imaging and Infra Red. Infra red imaging is greatly reduced by things such as rain, fog and/or smokke whereas Thermal imaginging will still work very well in those conditions. That's why the military prefers thermal imaging to infra red imagin on scopes etcetera. Thermal imaging can pick out troops lying in the snow or even underthe snow. Again, infra red and thermal imaging are two very different things.

If thermal imaging doesn't use the infrared part of the electromagnetic
spectrum, then what part do you suppose it uses?


--
- Frank Krygowski

I have infrared binoculars. Not thermal imaging. TI costs another 2500.

I'm not looking I went thru that 8 years ago. with the binocs I see emiited heat . The imager is more a receiver then computer to screen. Guess One

I'll look....

opinion is yawl following the lead of technoville when a key find device will start for 10000 cars.

these cars are toys. Tesla Toys. The fatal in Fla was a toy ride.

On Thursday, February 2, 2017 at 6:30:30 PM UTC-5, DATAKOLL MARINE RESEARCH wrote:
On Thursday, February 2, 2017 at 6:07:27 PM UTC-5, Frank Krygowski wrote:
On 2/2/2017 4:44 PM, Sir Ridesalot wrote:
On Thursday, February 2, 2017 at 3:11:58 PM UTC-5, Frank Krygowski wrote:
Snipped

I wonder how well those things work. They're said to use thermal
imaging, IOW infra red, I suppose. But ISTM that at certain ambient
tempreratures there might not be much temperature difference between the
bulk of a cyclist's image and the surrounding road. That would depend
on choice of outerwear. A jacket over a sweater, for example, might be
the same temperature as the air, especially in front where it's being
cooled by the relative wind. Helmets would certainly be at ambient
temperature on their outside. Cyclists with rain capes or with front
loads might have their legs hidden from such a camera.


There is a HUGE difference between Therma Imaging and Infra Red. Infra red imaging is greatly reduced by things such as rain, fog and/or smokke whereas Thermal imaginging will still work very well in those conditions. That's why the military prefers thermal imaging to infra red imagin on scopes etcetera. Thermal imaging can pick out troops lying in the snow or even underthe snow. Again, infra red and thermal imaging are two very different things.

If thermal imaging doesn't use the infrared part of the electromagnetic
spectrum, then what part do you suppose it uses?


--
- Frank Krygowski

I have infrared binoculars. Not thermal imaging. TI costs another 2500.

I'm not looking I went thru that 8 years ago. with the binocs I see emiited heat . The imager is more a receiver then computer to screen. Guess One

I'll look....

opinion is yawl following the lead of technoville when a key find device will start for 10000 cars.

these cars are toys. Tesla Toys. The fatal in Fla was a toy ride.

hmmmm hmmmm I remeber there was a distinction. 3-4 generations of equioment for sale specd the diff between infra and thermal imaging


these different types were off the market gone nnnn 7=8 years ago


http://www.opticsplanet.com/s/best-t...maging-hunting

there were , for sale, 3-4 generations of infrared and these g

On Thu, 2 Feb 2017 15:11:55 -0500, Frank Krygowski
wrote:

I wonder how well those things work.

Probably rather badly since they were intended for other purposes. It
will take some time and effort to turn them into a people detector and
to get the pattern recognition software right.

They're said to use thermal
imaging, IOW infra red, I suppose.

As Sir Ridesalot mentioned, there are different flavors of infrared.
In this case, it's short wave IR, which is what an IR illuminator,
some night vision adapters, the common PIR motion detector all use, TV
remote controls, and which most digital cameras can display. The big
advantage of short wave IR is that it can use cheap silicon sensors
and plastic optics.

On the other end of the IR spectrum are the long wave IR heat
sensitive (thermal) detectors. These usually need to be cyrogenically
cooled, and require expensive germanium lenses. This is what FLIR is
selling:
http://www.flir.com
and what is inside the camera I mentioned. It has no problem
recognizing any heat source (including the hot motor used in
mechanical doping). The rest is plotting the outline of the warm
object along the highest contrast boundaries, and feeding that to
pattern recongition software. It can certainly recognize a human, but
I have my doubts if it can even see a bicycle, which will tend to be
at ambient temperature. So, for the bicycle, it might switch to
visible light and uses the same contrast boundary and pattern
recognition system. It doesn't need to be very sophistocated. For
example, it could look for 2 wheels and ignore the rest as most (not
all) bicycles have 2 wheels.

Note the different parts of the IR spectrum:
https://en.wikipedia.org/wiki/Infrared#Commonly_used_sub-division_scheme

But ISTM that at certain ambient
tempreratures there might not be much temperature difference between the
bulk of a cyclist's image and the surrounding road.

Things look very different at different optical wavelengths. This is
the theory behind the image enhancement machines used to extract faded
writing from historical documents and read weathered writing on
stones. The machine illuminates the object with a variety of
different optical frequencies (colors) from IR to UV, and combines the
results. At some frequency (color), the contrast will be usable.

Our porch has a light controlled by a motion detector, but it too seems
to be IR based. There are times it refuses to respond. I assume that's
due to lack of temperature contrast.

Nope. Due to lack of correct lateral motion.
https://learn.adafruit.com/pir-passive-infrared-proximity-motion-sensor/how-pirs-work
In the common RIP detector, the contrast is produced by the Fresnel
lens, which is then detected producing an AC (alternating current)
output, which triggers the switch. If the PIR detector is oriented
perpendicular to the direction of motion, it won't produce any output.
Also, some Fresnel optics are better than others, along with the usual
focus compromises between close and far detection.

If you don't like the far IR imaging method, it is also possible to
identify objects using LIDAR (light RADAR) as was used in the DARPA
driverless car challenge. It was a complex system, but it worked well
enough at detecting obstacles, but because it's probably the most
expensive part of the puzzle, there's some question as to whether it
will be used in driverless cars. Telsa uses RADAR and visible light
optical sensors, while Google uses LIDAR:
http://www.allaboutcircuits.com/news/tesla-vs-google-do-lidar-sensors-belong-in-autonomous-vehicles/


--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On 2/2/2017 10:05 PM, Jeff Liebermann wrote:
On Thu, 2 Feb 2017 15:11:55 -0500, Frank Krygowski wrote:



Our porch has a light controlled by a motion detector, but it too seems
to be IR based. There are times it refuses to respond. I assume that's
due to lack of temperature contrast.

Nope. Due to lack of correct lateral motion.
https://learn.adafruit.com/pir-passive-infrared-proximity-motion-sensor/how-pirs-work
In the common RIP detector, the contrast is produced by the Fresnel
lens, which is then detected producing an AC (alternating current)
output, which triggers the switch. If the PIR detector is oriented
perpendicular to the direction of motion, it won't produce any output.
Also, some Fresnel optics are better than others, along with the usual
focus compromises between close and far detection.

Ah. OK, I didn't know the details of how those things work. Thanks.


--
- Frank Krygowski

On 2/2/2017 6:28 PM, Sir Ridesalot wrote:
On Thursday, February 2, 2017 at 6:07:27 PM UTC-5, Frank Krygowski wrote:
On 2/2/2017 4:44 PM, Sir Ridesalot wrote:
On Thursday, February 2, 2017 at 3:11:58 PM UTC-5, Frank Krygowski wrote:
Snipped

I wonder how well those things work. They're said to use thermal
imaging, IOW infra red, I suppose. But ISTM that at certain ambient
tempreratures there might not be much temperature difference between the
bulk of a cyclist's image and the surrounding road. That would depend
on choice of outerwear. A jacket over a sweater, for example, might be
the same temperature as the air, especially in front where it's being
cooled by the relative wind. Helmets would certainly be at ambient
temperature on their outside. Cyclists with rain capes or with front
loads might have their legs hidden from such a camera.


There is a HUGE difference between Therma Imaging and Infra Red. Infra red imaging is greatly reduced by things such as rain, fog and/or smokke whereas Thermal imaginging will still work very well in those conditions. That's why the military prefers thermal imaging to infra red imagin on scopes etcetera. Thermal imaging can pick out troops lying in the snow or even underthe snow. Again, infra red and thermal imaging are two very different things.

If thermal imaging doesn't use the infrared part of the electromagnetic
spectrum, then what part do you suppose it uses?


--
- Frank Krygowski

Check out some videos that sow the difference between infra red and thermal imaging. here's one to get you started.

https://www.youtube.com/watch?v=IM0pt3wObJc

But both respond to _some_ of the infrared part of the spectrum.

The difference you seem to be focusing on is whether there's a separate
infrared "light" source to illuminate the object, or whether the system
is detecting only the infrared generated by the object. Thermal imaging
doesn't use a source. In at least some systems, there is an IR source.
That's the "illumination" part of the "IR illumination" portions of the
video you linked.

It's infrared in either case, even though the source (and thus the
wavelength) is different.


--
- Frank Krygowski

AAA...yawl find the bank robber ?

On Thu, 2 Feb 2017 20:43:32 -0800 (PST), DATAKOLL MARINE RESEARCH
wrote:

AAA...yawl find the bank robber ?

Yep.
http://www.santacruzsentinel.com/article/NE/20170131/NEWS/170139938

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Thu, 2 Feb 2017 15:21:57 -0800 (PST), DATAKOLL MARINE RESEARCH
wrote:

as an old woman with a shopping cart ?
10 potted palms in barrels
a child stroking a flyer ?
a deer
a rabbit
a cow
mooooo

Patience. A technology is not considered worthless if can't solve
every problem. Like speech recognition, pattern recognition will
improve with experience, time, and research grants. You start with
recognizing humans. You then add fixed obstacles, vehicles, and
traffic aids. Maybe later, animals, speed traps, and politicians.

An acquaintance is working on an augmented reality vision system that
will identify birds by species. I've seen the prototype and it's
rather impressive. The fuzzy logic works much like fingerprint
classification and identification.

Of course, if you don't like all this, the government can just issue
RFID transponder tags, which you wear while riding. The vehicle will
belch some RF, causing the tag to return a signal and ID number. The
range can be easily calculated. If too close, the brakes are
automagically applied.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Thu, 2 Feb 2017 15:30:27 -0800 (PST), DATAKOLL MARINE RESEARCH
wrote:

I have infrared binoculars. Not thermal imaging. TI costs another 2500.

First, 2nd, or 3rd generation? Or perhaps just an image intensifier
(star light scope)? It doesn't really matter because these only
respond to short wave infrared and will NOT respond to heat, which is
long wave.
https://en.wikipedia.org/wiki/Night_vision_device

Incidentally, I have several:
http://802.11junk.com/jeffl/pics/AN-SAR-4B%20night%20vision/
This one has some rather cool optics. It's made for piloting a boat
in the fog. I've used it to drive my car at night using IR filters on
my headlights. It would a really good get-away car. I suppose I
could mount it on a bicycle, but it's really big, really heavy, and
sucks lots of power.

I'm not looking I went thru that 8 years ago. with the binocs I
see emiited heat . The imager is more a receiver then computer
to screen. Guess One

Wrong. You're not looking at heat. Try a hot stove, hot frying pan,
or anything hot that does NOT emit light. You won't see anything with
your IR binoculars. Basically, if your IR viewer does not have a cryo
imager cooling system, it's not going to see heat.

Fun and games: Find a cheap digital camera and look at the IR emitted
from a TV remote control.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann writes:

On Thu, 2 Feb 2017 15:21:57 -0800 (PST), DATAKOLL MARINE RESEARCH
wrote:

as an old woman with a shopping cart ?
10 potted palms in barrels
a child stroking a flyer ?
a deer
a rabbit
a cow
mooooo

Patience. A technology is not considered worthless if can't solve
every problem. Like speech recognition, pattern recognition will
improve with experience, time, and research grants. You start with
recognizing humans. You then add fixed obstacles, vehicles, and
traffic aids. Maybe later, animals, speed traps, and politicians.

An acquaintance is working on an augmented reality vision system that
will identify birds by species. I've seen the prototype and it's
rather impressive. The fuzzy logic works much like fingerprint
classification and identification.

Of course, if you don't like all this, the government can just issue
RFID transponder tags, which you wear while riding. The vehicle will
belch some RF, causing the tag to return a signal and ID number. The
range can be easily calculated. If too close, the brakes are
automagically applied.

That'll teach those illegal aliens.

--