#robotics Logs

May 25 2015

#robotics Calendar


00:02 eadthem then it goes in to cool down at the fastist rate our ovens can handle 1.8 to 2.5C per sec (We want to be at 2 to 4C per sec cool down rate.
00:04 Tom_itx i don't have parts yet anyway.. i'll see how it goes
00:04 eadthem of corse that and the speed of the belt all change from profile to profile, we use standard profiles and predict what i a board needs based on weight, and component coverage (Its not as scientific as it sounds).
00:04 Tom_itx goin to sleep now
00:04 eadthem the cool part is we can send a board thrue with probes on it, mesure what it realy is
00:04 eadthem and then the computer will predict what would happen if we ran it with diffrent speeds and temps.
00:05 eadthem night
00:05 Tom_itx thanks
01:27 kllrd Hi all. Starting a robot project w/ ras pi. Can someone reccomend a good recharchagable battery system to power the pi and a powered usb hub? Planning on charging it via outlet and possible solar
01:57 Triffid_Hunter kllrd: folks tell me http://www.banggood.com/Original-XIAOMI-5V-2A-10400mAh-Power-Bank-For-Smartphone-p-923637.html is quite nice
02:12 rue_bed kllrd, there are usb battery packs
02:12 rue_bed look over ebay
02:23 kllrd ty ill check the link, and look into the usb packs
03:50 tws2172 hello
03:54 Jak_o_Shadows yo
04:01 tws2172 what are you guys up to?
04:02 deshipu um, nothing, really, just carry on, no taking over the world with killer robots here, not at all
04:03 tws2172 darn
04:05 tws2172 anyone here have experience with opencv? specifically haar classifiers?
04:07 shipu didn't know opencv had hair classif
04:08 tws2172 haar-wavelet based cascade classifiers :P
04:09 deshipu so they can't classify straight hair, I see
04:09 tws2172 i guess they could be used to classify hair
04:12 Jak_o_Shadows I haven't done anymore than edge detection
04:13 tws2172 haar classifiers are really cool. I use them on several of my robots. It's a great way to find and track specific objects
04:19 tws2172 The biggest issue is training them. The process itself can take days. I've found that if you spin up one of those droplet servers and train on those, it takes a matter of minutes.
04:20 Jak_o_Shadows You'd ahve to pull the data set in by hand though
04:20 Jak_o_Shadows or train it in stages
04:22 tws2172 yeah, the data set takes a while to transfer over. But it still beats days of training
15:14 eadthem morning
16:36 tws2172 hello