I’m currently playing with wavelets as a part of my journey to know something about signal processing. Edge detection is a now-classic application of wavelets, so I figured I should experiment with that a bit.
To that end, I went out looking for a webcam that works well under Linux. Based on reports that the camera works in Linux — mostly this review, I picked up the Logitech QuickCam Vision Pro for Mac. I’d like to report that the camera does, in fact, work via the UVC driver with Video 4 Linux 2. Even auto-focus works (I’ve heard that this camera implements it in hardware, though I cannot confirm this).
It turns out that OpenCV has its own V4L2 interface, and that this has been exported to the OpenCV Python bindings. I grabbed some sample code from here and it worked out-of-box. With a 1-line modification, I was applying the Python Image library’s
CONTOUR filter. Here’s the result, plus blurry because of my shaky hands:
Works great, almost in real time on my desktop machine. So now the task ahead of me is to implement the same functionality, since that’s what got me started down this path to begin with.
Here are two pictures I took of the camera after removing the stand it came with. To be clear, these were (obviously!) not shot on the camera itself.