Hacker News Reader

As someone who has spent the past month fighting with FLIR customer support to get a $5000 camera working robustly and correctly I doubt that this will be a good product. We have dozens of FLIR cameras and at least a quarter of them have been sent back for re-calibration. We're about to send one out for the second time. Also, forget about open source support.

View on HN

This is meant to be pretty basic. Can you think of anything that could replace it? It seemed to work well when I demoed it, but for pro stuff obviously you wouldn't want to rely on it.

The FLIR machine vision cameras for industrial applications are driving me nuts. To a degree all machine vision cameras can be persnikety when you want to try and use them with FOSS software like ROS. Right now we are fighting with non-uniform correction of the cameras. Basically this is the camera's auto exposure. It consistently gets all out of whack. Also the GigE IP stack on the camera was a righteous bitch to get working. Once you buy the camera you have to spend a small fortune for an GigE SDK that doesn't even work on anything but CentOS. We ended spending a whole weekend doing some basic kernel patches to get the damn thing to kinda work on Ubuntu.

All of this is to say, it will probably work, but don't plan on building after-market apps on top of the camera technology.

Reply on news.ycombinator.com