It had always intrigued me about the operation of an optical mouse.
I just assumed it used some optical targeting method much like a laser system would to detect movement how ever, that would seem to require components of a costly nature due to the proximity of the surface and mouse.
So I did some online research. My first hit was a nice PDF file that explained it as a video capturing sequence and then processes the images to determine the direction of movement from the textured surfaced, etc..
How ever, I have some doubts if all mice are creative totally in this fashion. For one thing, I would think the over head cost on these devices wouldn't be as cheap as they are today? But maybe I am wrong in that respect. But the biggest issues that boggles me is, what if the surface is perfect with no visual imperfections? how does this method detect movement?
I just put my optical mouse on a sheet of clean white plastic that even under my bench scope (12x mode), it seems to be clean how ever, the mouse has no issues at all with this.
My son had a mouse pad that had two different images depending on the viewing angle. When he went to an optical mouse it got totally confused by the little embedded plastic lenses that give the image dependent on angle effect. He had to go back to a plain mouse pad (or the desktop).
Some mice can be hacked as video cameras. By writing frames to the screen and swirling the mouse around, you can copy your mousepad to your screen. Of course, rotation and tracking screw it up so you get curved paths and bad alignment, but that's to expected.
No big deal, it's only a small camera. You could build one yourself using, oh, 256 SMT photodiodes, let's say, plus an FPGA with analog mux and ADC input to do the chugging. And a lens so the photodiodes act as proper camera, and a few LEDs to illuminate the subject. (Course, you'll end up with something bigger than the very first mouse with wheels, but nevermind that.) Monolithic integration is a lot smaller, suitable for mass production, and sure, spin a couple million of them and you'll be paying only a few bucks each.
What's more remarkable to me is getting the image matrix quite right. You need to rotate your reference frame, so it's rotation invariant. Some very particular linear algebra required here. On the plus side, given enough processing to handle it, you get a lot of data available for just two outputs (delta X and delta Y), and that allows some pretty impressive things, like subpixel accuracy and operation on remarkably flat surfaces.
If it's *perfect*, then it doesn't. Either it goes nuts thinking it sees motion when it doesn't (more of a divide-by-zero error), or it sees no change and therefore determines no motion at all.
The smallest difference may be enough to register movement. Consider if it uses an 8 bit ADC: the difference between 169 and 168 is impossible to tell by eye, but it's quite possible electronically.
In my experience, reflective surfaces make optical mice go bonkers. A well-worn plastic surface mousepad, or a well-worn formica table, is usually terrible for tracking. Wood grain is quite excellent. Right now I'm working on a semigloss varnished white oak desk and my old Microsoft mouse tracks just fine up to its refresh limit.
I wonder what the ideal surface is. Periodic grids are also terrible, though rarely encountered. Maybe an aperiodic grid would be good? Ah, but it should have a fairly regular period over the entire area. White noise should work. Maybe Penrose tiles would be good too:
On a sunny day (Mon, 12 Oct 2009 09:49:59 -0700 (PDT)) it happened Tim Williams wrote in :
My formica desk is perfect for my optical mouse. It is 30 years old or older, and what is well worn? Hardly a scratch! Here you can see the pattern the formica has: ftp://panteltje.com/pub/col_pic/RGB_LED_strip_img0906.jpg That mouse has been mousing over it for tenth of years....
Do you use sandpaper on your desk? I even soldered on it, build whole projects on it, no damage. No mousepads for me.
Last formica surface I used was a biege desk at the library. The mouse spot has clearly seen a lot of hands, being polished nicely, and the mouse used there doesn't handle it very well at all. I stretched my arm all the way out to use it on a more dull area. I didn't notice what the mouse was, but I do remember the computer was some black IBM box.
I have a green marble sort of surface on my Bench, but it's too caked with rosin and such to use with a mouse. And electronics projects. It doesn't stand up well under hot solder, either; the surface turns brown and bubbles up. Probably cheap stuff.
formatting link
Hmm, that's a particularly dishonest image, rather old. It's messier these days.
The only thing pegboard is good for -- keeps those cables out of the way when not in use. Tools have never hung from it, I insist on keeping them out, front and center. :-)
I'm currently using a piece of graph paper laminated between sheets of clear plastic, and it seems to work fine. The squares are about 2mm in size, and there's a thicker line every 5 squares, and a thicker line again every 10 squares. I guess the periodicity is large enough compared to its viewing area that it doesn't get confused.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.