SpiderGears's profile - activity

@berak : Have you worked with opencv 3.0 API...? Having some issues. Any means to contact you...?

yeah something looks like magic.... since most values are -16 as in image data... image should more like something plane than the texture here.

@berak Any idea abou this?

hey!! i am working with disparity maps to find distances to real world object. I get a decent disparity image... so printed to content of the disparity matrix... and most of the values like 80-90% are -16 Any idea why..? Attached are Disparity Image and Image Data Disparity Image Data

Also disparity maps here looks too faint to me... Any help on this will be appreciated.

manually have a look at the two frames to discover point of similarity and then set the two ROIs accordingly.

Hey!! I am trying to work Minoru cam with opencv. I intend to generte disparity maps with the camera capture. I get the following error... As per the design of Minoru both the cams connect to same USB port.

HIGHGUI ERROR: V4L/V4L2: VIDIOC_S_CROP

HIGHGUI ERROR: V4L/V4L2: VIDIOC_S_CROP

VIDIOC_STREAMON: No space left on device

Here is output of bandwidth allocation when using one cam with guvcview application

sudo cat /sys/kernel/debug/usb/devices | grep "B: "

B: Alloc= 0/800 us ( 0%), #Int= 0, #Iso= 0

B: Alloc=251/800 us (31%), #Int= 4, #Iso= 5

okay.... Working with guvcview and this cat command i fig out i need work with lower frame rate like 15 fps to support two cams on single bus.

Unfortunately opencv is not able to set properties on my camera....

Any other way to set camera properties for capture??

as per my experience with opencv, the cams need not only to be of same make, also should be from the manufacturing batch to get optimal results.

I had earlier two cameras that worked fine. Later one of them broke, so I got one of the same make, but this didn't work as per the expectations.

You can try Minoru 3D cam... I haven't tried them yet.. waiting for the order..

Any tutorials or any other relevant resources for learning an implementing multi-label classification using SVM in OpenCV.... ?

I intend to use SVM classifiers using Bag of Words method for object detection. Sample Source code will a great help.

Don't post Questions blindly... Opencv is OpenSource hence look into the source code. See the module where the function is defined (Use docs.opencv.org) Browse module source on Github or your local machine.

Here it is: Opencv Github Repo color.cpp

Pls. post relevant code. It helps.

If you are not using an actual stereo camera, I would recommend you to do calibration by taking the video, every time you use the setup. This is mostly because every small disturbance in your setup can cause large calibration errors, hence the resulting disparity maps and point clouds won't be good.

Also if you are using actual stereo camera, you can save the calibration matrices and load them on next use, saving you the trouble of calibrating again.

Hey I am trying calibrate a Stereo rig. Attached is the results i get...

Whats going wrong.. tried with upto 40 image pairs...

Code here

And in code i have stream to save the calibration matrices... but doesn't seem to work.
It creates just a empty file. 0 Bytes. Not able to fig out why.

Most of the code is from opencv samples.

yeah i am already working with waitKey(1)

I wanted get rid of it. I think it will a little cumbersome to do so... may i should ignore it.

okay... you want me do my own gui. Actually final application won't be gui based. So i guess i shouldnt be putting much time into it right now. Just wanted to explore options for debugging purpose.

thanks for the help though.

waitKey(0) means wait for keypress infinitely.

Not what I am looking for.

static boost::once_flag cv_thread_flag = BOOST_ONCE_INIT; boost::call_once(cv_thread_flag, &cv::startWindowThread);

found this is page here: http://web.ics.purdue.edu/~yuanl/2011/08/ros-cvnamedwindow-doesnt-show-up/

don't Understand it.

@berak: I am not using qt for my application. Since everything is on main thread... i think i own the gui.

I am truing to show the camfeed into a window using imshow() but it needs waitKey() after each frame. Referred to a few aritcles on web that say highgui requires waitkey() to be called in order to get time to update the window.

Is there a workaround with cpp interface to accomplish this without use of waitKey() ?

Well this a little complex problem and is an active area for research. For the fact I am working on the same problem. Here is the procedure I have adopted

1. Load your image containing text.

2. Now that you have your image loaded

   2 a) Extract pixels representing individual letters of text

   Approach 1: SWT (Stroke Width Transform) REF SWT needs to know if your text is Light-on-dark background or Dark-on-light background, if you do not know this then you have to make two passes of this algorithm.

   Approach 2: Flood fill your image repetitively until all pixel are covered. For every seed pixel mark the pixel that got included in the floodfill and do not run floodfill for these pixel again.

   For each seed pixel collected the relevant pixel and call the collection a component. Filter out non-relevant components using any appproach similar to the case of SWT.

3. Group the components to make text words, then to text lines and finally into text blocks.

4. Pass the text block to tesseract.

Current implementation of floodfill in opencv check here returns only the number of pixel marked for a seed pixel and not the pixel location. Hence to get pixel location in order form components some modification needs to be made. The marked pixel location can be collected in a Boost::unordered_map for better performance.

Also this is a github project that extracts text components from images using SWT approach.

first note: Avoid this.

This code basically had pointer of type unsigned character which is assigned the address of the yth row of the image. the address is cast to Unsigned char type before assignment.

for IplImage structure

->imageData is the start of image data skipping the image header data

->widthStep is the size of a row of pixels in the image.

right now i am working with 640X480

I am working an opencv Application on Linux plateform. It is to have multi-threaded architecture. Demand of processing is on a higher side: +Generating real time disparity maps from live video feed from two cams. +Edge detection and further processing. +Object Recognition and Detection in real time. +Face detection and recognition in real time.

All processing is to be done on the same video feed in real time on threads in parallel.

I want to move this application on to a portable processing board in order to get a highly portable system that is really compact and light. (like a medium sized hand held device)

What board should i go for ?

PS: I do not have any knowledge of working with any these boards.

remove jpeg support... i don't think it the solution or even right approach to this. this should have smthing to do with the parameters to traincascade though i am not able to figure out the exact thing.

we have tried training on Ubuntu 12.10, 13.10 and Fedora 19. Got same message on all systems though at different stages of training. The message does not stop the training itself... Training continues but now consuming much less cpu resources than before (earlier it was using 100% of all my 4 cpus, after this jpeg msg it came down to 90% of just one cpu.. other cpus were left idle.)

I don't things any of the images are corrupt since training i do not get any such errors in early stages of training (all went fine up till stage 10 for me on Fedora machine..).

I am using opencv 2.4.7 downloaded form Sourceforge.net

Additionally you might encounter problems when using two cameras simultaneously specially on a USB 2.0 interface. The first camera connected to USB host tries to reserve all the available bandwidth leaving nothing for the second cam. Error Message: VIDIOC_STREAMON: No space left on device

Solution:

1. Calculate bandwidth required manually and make appropriate configurations.
2. Put cameras on different USB hosts.

Ref thread http://answers.opencv.org/question/4909/using-2-webcams-simultaneously/

Latest source can be obtained from a. Github Repository: https://github.com/Itseez/opencv b. Zip can also be downloaded from Sourceforge: http://sourceforge.net/projects/opencvlibrary

Add option WITH_TBB=ON when building opencv.

Following the instruction form the blog on opencv_traincascade... Coding-Robin. At Stage 9 of training i am this message

Premature end of JPEG file

What is it about?

Here is the command

opencv_traincascade -data classifier -vec samples.vec -bg negatives.txt -numStages 20 -minHitRate 0.999 -maxFalseAlarmRate 0.5 -numPos 1000 -numNeg 600 -w 60 -h 60 -mode ALL -precalcValBufSize 1024 -precalcIdxBufSize 1024 -featureType LBP

thanks @berak, now i understand your point.

Hey!! I am new to OpenCV. Want to implement stereo vision so i bought two webcams from iball this. These do not seem to have autofocus feature. Does this affect the performance and results.

a) What works better cams with autofocus or without..?

b) I would also also like to the same camera feed for object detection and recognition in view of upto 5m atleast. How does this affect the results.

c)Any other cam that you may suggest for the same.

All tasks to be performed in real time.

Once all set I want to move the system to an embedded board such as Beagle Bone or something similar. Once project repository is setup.. any contribution from the community is welcome and will be hugely appreciated.

Thanks the beautiful insight. I will am little more interested in your final results like the accuracy obtained and the performance of the system. Did you process the video feed on a Computer/Laptop or an embedded board such as a Beagle Bone or something.

If a project repository maintained that be really helpful.