keltik's profile - activity

There are Step-by-Step instructions provided here.

Hello, Are there any Step-by-Step Tutorials that make openCV functions availible from within a "Blackberry-Cascades-Project" created inside the Momentics IDE on a Windows 7 64-bit machine?

I already found HelloOpenCV, but i do not have "make" and "qcc" on Windows 7 64-bit. I tried to install make and then the Momentics IDE showed me an error "cannot find qcc". I know that it is something like a compiler...

Secondly, i also found Blackberry OpenCV, but there is no explanation on how to integrate everything within the Momentics IDE and a Cascades-project.

I am used to getting things done in Java and i do understand the concept of importing libraries from the Java perspective.

I found a Visual Studio Tutorial and wanted to ask if it is possible to use this to configure the Momentics IDE?

I am writing a scientific document and i have cited a website from the opencv documentation. The website used in the cite is:

http://opencv.willowgarage.com/documentation/cpp/saturation_arithmetics.html

Does someone know the author ?

found the error: i had a typo in path. question can be closed

downvote?! give reason please. I will correct the issues.

I want to calculate the running average of 2 pictures (Later i am going to add more). I am using java opencv. i have written this piece of code, which throws an error in c++ code. I cannot understand why. Could someone clarify??

import org.opencv.core.Core;
import org.opencv.core.CvType;
import org.opencv.core.Mat;
import org.opencv.highgui.Highgui;
import org.opencv.imgproc.Imgproc;

public class BackgroundAverageTest {
    public static void main(String[] args) {
    System.loadLibrary(Core.NATIVE_LIBRARY_NAME);
    Mat pic1 = Highgui.imread(pathToMyPic1);
    Mat pic2 = Highgui.imread(pathToMyPic2);
    Mat accumulated = new Mat(CvType.CV_32FC3);
    Imgproc.accumulate(pic1, accumulated);
    }
}

Here is the Exception:

Exception in thread "main" java.lang.Exception: Unknown exception in JNI code {imgproc::accumulate_11()}
    at org.opencv.imgproc.Imgproc.accumulate_1(Native Method)
    at org.opencv.imgproc.Imgproc.accumulate(Imgproc.java:1701)
    at tests.BackgroundAverageTest.main(BackgroundAverageTest.java:17)

As you mentioned already: It is probably due to the byte cast. Could you clarify why you use such a tricky approach to set and get values ?

You can check out this resource:

http://www.codeproject.com/Articles/10248/Motion-Detection-Algorithms .

Although its made for python, i guess it is possible to understand the general technique and adapt to some other language.

As long as your background is fixed, you can track any object with it.

I am using OpenCV in Java. Is there a method that destroys a Mat object (deletes it from memory)?

the link lacks the step with "Native library location" and "add c++ dll", it also confused me when starting to program openCV in java.

check out this post for the answer: http://answers.opencv.org/question/13450/please-help-me-in-configure-the-opencv-for-eclipse/ (dont forget to upvote the post)

If my solution is working for you: i would appreciate it you accepted it as a solution by clicking the accept sign.

Which method is faster for retrieving a submat, adjustROI() or submat() ? I want to know it because im working with java and i got a loop over many images of size 1600*1200.

Thanks in advance.

Although Mr. Pavlenko was clear and has answered the question, I want to add something, namely using this code:

FeatureDetector fd = FeatureDetector.create(FeatureDetector.FAST);
...
fd.detect(myImageList,myImageKeyPointList);

If myImageKeyPointList is empty also an exception is raised (i guess that is the same reason as Mr. Pavlenko mentioned in his post):

Exception in thread "main" java.lang.IllegalArgumentException: Incomatible Mat
    at org.opencv.core.MatOfKeyPoint.<init>(MatOfKeyPoint.java:31)
    at org.opencv.utils.Converters.Mat_to_vector_vector_KeyPoint(Converters.java:591)
    at org.opencv.features2d.FeatureDetector.detect(FeatureDetector.java:229)
    ...

Besides i am using openCV 2.4.5 and java.

You must supply the c++ dll AND the openCV-jar to the jar file at jar creation time. Although i am not sure how to do it, i am sure about that this is the error, because the same error happens when forgetting to add the dll to the classpath.

You have to do this (Step by Step explanation):
1.) Start eclipse, open new Project
2.) right click project, go into "properties"
3.) go into "Java Build path", click add external jars
4.) go into the folder where your jar is placed, "open"
5.) select the newly added jar and expand its "plus-sign"
6.) select "Native library location" and click edit
7.) here select the location where you have stored your c++ dll

If you do not know where your c++ dll and your jar file is:
-download openCV
-unpack it go into folder "build/java" (you will see opencv-245.jar)
-go into "build/java/x86" (you will see your dll)

I want to read a fixed-size-submat from a bigger Mat while sliding it over the bigger one pixel by pixel (or maybe 5 pixel by 5 pixel or somtheing like that, i hope you get the idea). Is there a predefined function for that in openCV, if so which one is it? (Besides i am using Java with opencv 2.4.5..) I have an idea how to write it myself in java using Mat.submat(a,b,c,d), but i thought that the underlying c++ dll implementation would be a lot faster.

Thanks for the side-note that it should be easy to adapt.

I have read the Tutorial on calculating and drawing histograms in c++. Unfortunately, i couldn't find the "histImage()"-method inside the Java API. Do I have to write a Histogram-function from scratch or has somebody experience with java openCV and could point me in the right direction respectively point to the corresponding methods in the java API.

yep i found it out

Why does the png-file-size increase? Here is the code (in Java):

Mat m = Highgui.imread("C:/img1.png");
Highgui.imwrite("C:/img1out.png", m);

Initially the "img1.png"-file is 900kb, but "img1out.png" becomes 2200kb! I have read that the compression in openCV is not as good as in other image-processing API's, but recently some compression has been added.

I tried out to do this here:

Mat m = Highgui.imread("C:/img1.png");
MatOfInt mPngCompressionRate = new MatOfInt(9);
Highgui.imwrite("C:/img1out.png", m, mPngCompressionRate);

But the size still goes up to 2200kb of "img1out.png". What am i doing wrong here ?

Actually i had to instantiate the "compression-vector" like this "MatOfInt mPngCompressionRate = new MatOfInt(Highgui.CV_IMWRITE_PNG_COMPRESSION,0);" gives 5000kb and "MatOfInt mPngCompressionRate = new MatOfInt(Highgui.CV_IMWRITE_PNG_COMPRESSION,9); gives 2000kb". (poor javadoc)

totally forgot about those pages, i hope they will help, i am going to post the result here, thank you