Re: infrared beacon for pet door
- From: Jan Panteltje <pNaonStpealmtje@xxxxxxxxx>
- Date: Sat, 13 Dec 2008 13:22:29 GMT
On a sunny day (Fri, 12 Dec 2008 14:45:52 -0500) it happened "Stephen D.
Barnes" <stephendbarnes@xxxxxxxxxx> wrote in
<IpednXY1GYDCIt_UnZ2dnUVZ_judnZ2d@xxxxxxxxxx>:
Ok, I just did some coding, and Iwas thinking like this:
They claim the 'curve' is unique (say outline) for the 'correct' cat.
I did a full FFT open the pictures from my security cam,
hey a 640x480 picture in color, really, it is a lot of data.
The idea being that the spectrum should be unique, but
shift to higher frequencies for a further away object.
However the ratio between the FFT components should not change much
when the object is closer or further away?
I am not sure if that would hold during rotation and sizing due to distance.
It is a lot of data per picture.... result is 230408 abs frequency values....
The soft is fast enough it seems.
Anybody any ideas?
Would almost think you could use a vactor descrition, like in characters in
a pdf document, to store outlines.
image to vector graphics?
Have to think about this some more, any suggestions welcome.
.
- References:
- infrared beacon for pet door
- From: acannell
- Re: infrared beacon for pet door
- From: GregS
- Re: infrared beacon for pet door
- From: Jim Thompson
- Re: infrared beacon for pet door
- From: Joel Koltner
- Re: infrared beacon for pet door
- From: David Lesher
- Re: infrared beacon for pet door
- From: Stephen D. Barnes
- infrared beacon for pet door
- Prev by Date: Re: Digitally-controlled pot?
- Next by Date: Re: radio tower in Likely, CA
- Previous by thread: Re: infrared beacon for pet door
- Next by thread: Re: infrared beacon for pet door
- Index(es):
Relevant Pages
|