chriszuma's profile - activity

Seriously? Is there no answer to this? I really don't want to change imaging libraries at this point over something so trivial.

Come on. Anyone?

This seems like it should be very simple, but I can't find any explanation of how to do it. I tried these and neither worked:

cv2.setWindowProperty("display",cv.WND_PROP_FULLSCREEN, cv.WINDOW_FULLSCREEN)
cv2.setWindowProperty("display",cv2.WND_PROP_FULLSCREEN, cv2.WINDOW_FULLSCREEN)

I just keep getting this error message: AttributeError: 'module' object has no attribute 'WND_PROP_FULLSCREEN'

I'm working on a real-time video processing application, and I need as little latency as possible. I'm suspicious that the standard namedWindow and waitKey solution is adding delay, because there is about 100ms of lag when it should be closer to 30. Also it is a bit superfluous because my application will be run on a system with no keyboard. It seems to take a lot of CPU power to display too, which I would like to cut down if possible.

So is there a faster/less CPU-intensive way to display my video feed? Something like DirectX, but on Linux?

Thanks.

How does one actually add SSE or NEON optimization? Do I have to recompile OpenCV from source, or do something to my code?

I know this is kind of a stretch because VideoWriter is so basic, but I would like to use the MKV (Matroska) video container format. It is nicer for our application because each frame is timestamped, and our framerate is not always constant. The AVI or MPEG files like to assume a constant framerate so the video speeds up and slows down if it's not encoded at exactly that framerate.

Does anyone know of a way to make this happen? Or if there's a better solution to my problem?

Thanks

I would advocate using whatever language you are most comfortable with. I'm partial to Python because it takes care of a lot of the details for you. Library calls will be the same speed.

So is NEON or SSE better? All I'm really doing is essentially some resizing, thresholding and blob detecting based on an AOI from an external sensor. It doesn't seem like I should need an i7.

Halcon is really slower? That's surprising considering they charge $700 + $200/PC for their software. I guess I'm glad I asked before I wasted a bunch of time trying to get it to work.

I'm pretty new to computer vision, but I am attempting to do some real-time video overlay stuff that is relatively simple. I've implemented a nice version that runs on saved videos and I'm ready to step up to our end goal of having a really nice machine vision quality camera integrated into the system.

To that end I acquired a GigE IDS uEye 5240CP and have been trying to shoehorn it into my OpenCV application with very little success. It turns out the camera doesn't use a standard communication protocol so I am messing with the pre-alpha version of an open-source python wrapper for their API. There has to be a better way.

So my question is what cameras are easiest to integrate into OpenCV? My main requirements are HD resolution, global shutter, 30fps, very low latency, and compatibility with Linux (and of course OpenCV).

Any advice in the right direction would be deeply appreciated. Thanks.

Hi all,

I've been developing a real-time augmented-reality type application in Python OpenCV, and had a rather nasty surprise when I switched from my Core i7 development machine to the Atom unit that we wanted to actually use. It turned out to be nowhere near fast enough, giving me 5-10 fps where it would easily keep up with 30 fps videos on the i7.

So my question is whether porting my application to C would help at all, or if it's just that OpenCV can't handle this kind of application. When I put my app through the Python profiler almost all of the time was being spent in cv calls, and waitKey was also taking forever. Should I just give up and look at something like Halcon?

Thanks,
- Chris