As I mentioned, I have some cool stuff to share. The first is I bought a new dashcam. I was inspired by this extraordinary Youtube channel which shows an obsessively well-organized endless stream of car crash videos as seen from the perspective of the cameras that almost all Russians use when driving. If you don’t think that self-driving cars are one of the most important goals of technological progress for the first world, you can just watch that channel until you’ve changed your mind.

The camera I bought was only $58, so it might not last more than a few weeks (it’s worked fine now for 2!). The image quality is excellent, night and day. I’m quite happy with it. It doesn’t have a battery which I thought was wise given that my car often reaches oven-like temperatures. If the thing lasts, I’ll hard wire it to the cabin lights.

That’s part one. The next thing I can show off is some cool OpenCV work. The first project was to find, roughly, lane lines. I thought I’d try my algorithm on my own video footage. Here’s a pretty interesting video showing a lot of noteworthy things.

I found it eerily calming to safely watch a drive I do way too often. It’s like the difference between watching a horse and being dragged behind one. The video starts just after what I believe to be the most dangerous on-ramp in the world. Having survived that, we can check out the camera, my driving, and my algorithm’s ability to find lane lines. Note that this is downsampled from full HD (1920x1080) from the camera to quarter HD (960x540).

You can see that the algorithm sometimes has trouble with bridge shadows. Parallel stress cracks can be tricky too, but that’s true for my human driving abilities too! It’s best not to obsess too much over this algorithm since more work is planned for it. Let’s just say that knowing where the lanes are is not a trivial problem for a computer. I’m not quite ready to let my car pilot its own course just yet.