|
Re: Blob Analysis: msg#00233lib.opencv
--- In OpenCV@xxxxxxxxxxxxxxx, "rickypetit" <rborras@...> wrote: > > --- In OpenCV@xxxxxxxxxxxxxxx, "ivanhan83" <exxonvald@> wrote: > > > > Hi everyone, > > > > Using CvCam and after some filtering, I've obtained a binary image > > (black BG, white blobs). How do I go about identifying these blobs? > > Basically, I just want to bound each blob with a rectangle. What's an > > effective way to do it? I've tried cvFindContour, the blob library, > > etc, and they all seem to take quite a bit of processing time (I could > > also be implementing it wrongly). I've even tried flood-filling (i.e. > > identify a white pixel and do BFS until you get all the blob, but > > that's slow too). Would really appreciate it if someone could point me > > in the right direction. Thanks a bunch! > > > > Ivan > > > How many images per second do you have to process? Using the blob > library we can process up to 15 images per second (12 of 640x480 and 3 > of 1024 * 640 ) (extracting blobs and doing many other things ). We > use the library in release version, compiled using the Intel Compiler > (v8.0) (the performance increases a lot using Intel's compiler). > > Hope it helps, > > Ricard > Which function to use if we want to filter small blobs in that blobslib?Can we modify or add some new functions to the blobslib?The comments for the functions are in french,can't very understand it.-_-! Change settings: http://www.yahoogroups.com/mygroups, select Get Emails (get all posts) Daily Digest (one summary email per day) Read on the web (read posts on the web only)Or Unsubscribe by mailing OpenCV-unsubscribe@xxxxxxxxxxxxxxx |
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| Previous by Date: | Re: OpenCV project (dis)organization: 00233, Olivier Bornet |
|---|---|
| Next by Date: | Opencv with libdc1394: 00233, andrehochuli |
| Previous by Thread: | Re: Blob Analysisi: 00233, rickypetit |
| Next by Thread: | Re: Blob Analysis: 00233, rickypetit |
| Indexes: | [Date] [Thread] [Top] [All Lists] |
| News | FAQ | advertise |