----- Original Message ----- From: "Erik Reuter" <[EMAIL PROTECTED]> To: "Killer Bs Discussion" <[EMAIL PROTECTED]> Sent: Monday, September 27, 2004 3:25 PM Subject: Re: Strike Zone
> On Mon, Sep 27, 2004 at 03:22:35PM -0500, Dan Minette wrote: > > > With two cameras following the baseball, this should be > > straightforward...just find the intersection point of two vectors in > > space. Multiple pairs of vectors can be taken, and continuity can be > > expected. I was usuing technology similar to this 25 years ago to > > work with bubble chamber photos. > > We're not communicating. Tracking the baseball position in space does > not seem difficult. The problem I am asking about is where the top and > bottom of the strike zone are for each individual player. How does the > system you are referring to remove the human judgement call on that? I quoted the article earlier: "Two cameras sit low and close to the field, just off the first and third baselines. They determine the batter's strike zone." There may actually be a human that clicks on a cursor twice, I'm not sure. The letters should be a no-brainer for automation. Knees can probably be recognized automatically...the pattern is fairly straightforward, I think. If not, then a click to get the knees is not all that hard. This is something for which there can be a lot of samples and tests before the actual event. Just get a lot of folks in every baseball uniform to walk out and have the system guess the height of the letters and their knees. I do a different type of pattern recognition, but this seems pretty straightforward to me. Dan M. "Two cameras sit low and close to the field, just off the first and third baselines. They determine the batter's strike zone." from a couple of posts ago. Dan M. _______________________________________________ http://www.mccmedia.com/mailman/listinfo/brin-l
