Move video file saves from detectcone to another service

[arbulgazar]

As of now, the video files are saved in detectcone This needs to be moved to another service to be compatible with opendlv-vehicle-view (possibly generated by opendlv-vehicle-view itself) as it preserves correct time-stamping and .rec replays.

• [ ] Remove .png save from detectcone
• [ ] Add opendlv-vehicle-view to started microservices
• [ ] Save .pngs in the opendlv-vehicle-view

[chrberger]

@arbulgazar The steps are much simpler: a) remove https://github.com/chalmersfsd/cfsd18-perception-detectcone/blob/efbea17045d34a0b13cb2dca2717431eb93f8d86/src/detectcone.cpp#L684

b) modify the yml file that you use for your data recordings and ensure that youhave: 1) a proper opendlv-device-camera-? service matching your camera that is producing I420 frames living in a shared memory environment that is configured correctly (width, height, etc.) 2) opendlv-video-h264-encoder to "connect" to the shared memory segment to convert the I420 frames into opendlv.proxy.ImageReading (also configured correctly) that can be later either transformed into .png files or restored into shared memory segments for offline testing

[arbulgazar]

I will remove the image saving in the detectcone.cpp We have a opendlv-device-camera-opencv starting in the testcases, so I guess the 1. is fulfilled. All we need to do is step 2. then?

[chrberger]

Yes, that's correct. And a dry-run in the lab of course.