Got Questions? Get Answers.
Discover MakerZone

MATLAB and Simulink resources for Arduino, LEGO, and Raspberry Pi

Learn more

Discover what MATLAB® can do for your career.

Opportunities for recent engineering grads.

Apply Today

Thread Subject:
Image Processing

Subject: Image Processing

From: Rohit

Date: 26 Aug, 2010 20:10:05

Message: 1 of 7

Hi, I m doing a project " wiper control of a car using image processing" .I need some help regarding ,how to detect rain drops and count the number?

Subject: Image Processing

From: Sean

Date: 26 Aug, 2010 20:51:26

Message: 2 of 7

"Rohit " <rohitronge29@gmail.com> wrote in message <i56hmt$gk3$1@fred.mathworks.com>...
> Hi, I m doing a project " wiper control of a car using image processing" .I need some help regarding ,how to detect rain drops and count the number?

I'll send you working code for this if and only if it's compatible with my 2003 Toyota Tacoma and totally complementary.

Subject: Image Processing

From: Walter Roberson

Date: 26 Aug, 2010 20:50:24

Message: 3 of 7

On 10-08-26 03:10 PM, Rohit wrote:
> Hi, I m doing a project " wiper control of a car using image processing"
> .I need some help regarding ,how to detect rain drops and count the number?

Larger raindrops can be detected with an array of force sensors. It might,
however, be necessary to do some detrending to remove the force effects of
gusts of wind, but in doing so you would need to take into account that the
wind would not press equally hard on all parts of the windshield. Another
difficulty you might have with the force-measurement approach is that the
front of a car will in effect act like a "wing", and at anything more than a
trivial velocity the air flow patterns over a wing-like-object become chaotic,
so force sensors on the windshield are going to be receiving inherently noisy
readings.


When the raindrops are sufficiently small, you start getting in to semantic
arguments about what exactly a raindrop *is*. Consider, for example, the
difficulty of counting the number of "raindrops" in a moderate fog.


But you mentioned "image processing". What kind of resolution have you got,
and what frame rate? Do you have some sample movies you could post on a web site?

Subject: Image Processing

From: Sean

Date: 26 Aug, 2010 21:02:05

Message: 4 of 7

Walter Roberson <roberson@hushmail.com> wrote in message
>
> When the raindrops are sufficiently small, you start getting in to semantic
> arguments about what exactly a raindrop *is*. Consider, for example, the
> difficulty of counting the number of "raindrops" in a moderate fog.
>
>
> But you mentioned "image processing". What kind of resolution have you got,
> and what frame rate? Do you have some sample movies you could post on a web site?

It has to be very stable too: if a small bug lands on your windshield, at highway speeds, it might appear as a rain drop. Thus if the pressure of the windshield wiper squeegee isn't quite enough to remove the bug (which in my experience it never is) it could find itself in a positive feedback loop which could burn out the windshield wiper motor. Then what's going to happen? It'll clearly have to display some type of msgbox with a few expletives and/or throttle the car since it's very unsafe to drive a vehicle with a dirty windshield.

Subject: Image Processing

From: Walter Roberson

Date: 26 Aug, 2010 21:39:23

Message: 5 of 7

On 10-08-26 04:02 PM, Sean wrote:

> It has to be very stable too: if a small bug lands on your windshield,
> at highway speeds, it might appear as a rain drop.

I was musing about the problem of bugs as well. In some areas at some times of
year, the bugs are extremely thick; I am told, for example, that "shad fly
season" (early June) is difficult driving near North Bay, Ontario, Canada.

Then there are bird droppings.

In most parts of Canada, and the more northern parts of USA, it can be random
as to whether any particular snow-flake melts just before hitting the
windshield, or melts upon hitting the windshield (the heat of impact pushing
past the melting point), or sits on the windshield and melts (due to heat
inside the car), or sits on the windshield and freezes due to the cold wind
blowing away the internal warmth that the snow had.

More difficult still for the algorithm might be freezing rain: cloud internal
friction or localized warm air can end up keeping rain droplets liquid even if
the average air temperature is much lower. When such a droplet hits something,
then especially if there is a high wind, the droplet may freeze near
immediately or very quickly -- faster than a normal "fast" cycle for wipers.
This can form a layer of ice, and any particular further drop might make it
_through_ the layer, or might freeze on _top_ of the layer, or might be liquid
enough to be pushed off the windshield by the pressure of the wind -- but for
the problem original wording, only the drops that actually make it through to
the windshield at some {delayed} point are to be counted. Liquid drops running
_up_ the windshield (due to wind from the front), chilling and freezing more
as they go up, are a common problem in winter driving; for the purposes of the
original request, this is a difficulty because the timing of the _hitting_ of
the windshield might be somewhat decoupled from the timing of the need for the
wipers.


A further difficulty in this part of the world is that when spring finally
approaches and there is a sunny day, snow melt quickly accumulates in puddles
on the road. This snow melt is typically *quite* dirty. And as cars travel
through it, the snow melt splashes onto nearby cars and windshields (and
pedestrians). The detected raindrop count for these events would typically be
relatively low, as they tend to be somewhat similar to being hit with a
raindrop with a volume of about 1 litre. Even with that small drop count, the
driver's view would be greatly blocked, and the important factor would not be
a short interval between wipes but rather a short latency; indeed considering
momentum issues, it would be best to detect these events *before* the water
hits and start moving the wipers so that they will be ready in time.


 > it could find itself in a positive feedback loop which could burn out
 > the windshield wiper motor

Or around here, it could find itself in a situation where the wiper was frozen
in place, and could burn out the motor trying to move them...

Subject: Image Processing

From: Ashish Uthama

Date: 27 Aug, 2010 12:00:02

Message: 6 of 7

What kind of raw data are you planning to acquire?

http://auto.howstuffworks.com/wiper4.htm

Subject: Image Processing

From: Rohit

Date: 11 Apr, 2011 10:12:04

Message: 7 of 7

"Sean " <sean.dewolski@nospamplease.umit.maine.edu> wrote in message <i56k4e$naf$1@fred.mathworks.com>...
> "Rohit " <rohitronge29@gmail.com> wrote in message <i56hmt$gk3$1@fred.mathworks.com>...
> > Hi, I m doing a project " wiper control of a car using image processing" .I need some help regarding ,how to detect rain drops and count the number?
>
> I'll send you working code for this if and only if it's compatible with my 2003 Toyota Tacoma and totally complementary.
>>please send me matlab code for it..

Tags for this Thread

What are tags?

A tag is like a keyword or category label associated with each thread. Tags make it easier for you to find threads of interest.

Anyone can tag a thread. Tags are public and visible to everyone.

Contact us