basheer's profile - activity

According to this SO answer, the OpenCV FitLine function already does this.

is it a missing feature or just bad naming of variables or parameters?

hmmm... interesting... are you sure everything worked? Did u compile stuff that use gpu module? Oh, well. I probably messed something up back then (I'm working on Ubuntu now and it works easier, I think)

I spent a couple of days trying to do the same on my macbook in OS 10.8... I ended up installing Ubuntu and doing it from there (was much easier with less bugs and errors). Good luck!

Why does the cv::gpu::HoughLinesP method not have a threshold parameter in it? What do I do if I wanted something like a threshold parameter? Is the maxLines parameter effectively the same?

BTW I'm using the most recent master branch version of OpenCV...

I was just wondering whether the canny() method in openCV applies a gaussian blur or not (I want to know whether I should use a gaussian blur before applying canny). Also, if it does, what parameter does it use by default for the blur?

It doesn't mention blur in the docs. However, the canny example here applies blur before applying canny(). So, I'm guessing the answer is no. I just wanted to make sure.

I went through the source code for canny(), but it doesn't explicitly call gaussian blur. I was thinking maybe it manually applies it (but I can't really read the source and understand it)...

I'm using openCV 2.4.5 btw

So, I recently found out (from here) that the new Java bindings for openCV (as of 2.4.4) don't include the gpu module.

How would I go about generating the modules myself? Can someone go over how this would be done (a summary or something would be fine)? Links, hints, anything...

Also, is this currently being worked on by the openCV team?

nevermind my old comment. I was confusing this with another question...

also, for the record, I found that using Oracle JDK 7 is better than 6 (less errors to deal with)...

@Andrey I solved that specific problem by fixing a bug that I reported here.

Thanks for your answer. Is there any way I can use the gpu module using the Java bindings? Would I have to write my own JNI stuff? Is there anyone currently working on this (perhaps I could help instead of working alone)?

Background: I've spent days figuring out ways to fix errors while compiling openCV in the new java bindings thing (I'm on 32-bit Ubuntu 10.04). I finally managed to get openCV 2.4.9 to compile using cmake (with no errors) with all the needed modules (java, gpu, python etc.) and with CUDA enabled. After the whole thing builds (again, no errors), I create an Eclipse project (all according to the tutorial here), and I can use the core modules and others (the example code runs fine). HOWEVER, I can't use the gpu module at all (it doesn't show up when I try to import it).

Summary: Once again, I used the WITH_CUDA=ON flag when I did cmake (and I checked the output, and it did actually build and create folders and .o and .a files for gpu stuff, but it never created the .java files for the gpu module, and so I can't access the gpu module in Java).

So, my question is: Is it even possible to use the gpu module using the new Java bindings? Or is there some sort of limitation with this new Java bindings thing? (I really have no idea) If it is possible, then I must be doing something wrong (I might open another question to solve that). But I'd first like to know whether it is possible or not.

I'm still having the same problem, although I have Oracle JDK 6, Apache Ant 1.8, and Python 2.6. Of course, I made sure that cmake actually SEES them all and recognizes them (I exported the appropriate variables) I'm running 32-bit version of Ubuntu 10.04

Are there any other dependencies I should worry about?