koraxis's profile - activity

Thank you for the suggestions. But the problem is, the camera preview is created by Vuforia , on JNI side, Not common Android Java side Camera. So i guess its not that optimised for speed that much

I am using OpenCV,Vuforia and Tesseract(tess-two) for my app. My system works like this:

1. Vuforia detects the target and sends the full frame to OpenCV (JNI)
2. OpenCV gets the frame and does some image processing to make it readable for Tesseract (JNI)
3. Tesseract(tess-two) takes the byte arrays from opencv and does the OCR processings on the image(byte-array)

I have problem with the 2nd and the 3rd parts.But since this is an OpenCV forum, i think i should ask the 2nd one. I am doing the image processing parts, as its written above, in C++ and the problem is , since i have to do image processing on every detected frame from Vuforia i create lots of image processing load.

What i want to do is, find a multithread solution for image processing so that the live camera preview that is on the screen wont have lags and will be smoother. Where should i start from for making the image processing faster ?

But type function returns the values for like CV_8U, or CV_16U etc. I was asking about getting the format like, if its RGB, BGR, GRAYSCALE, ARGB etc

How can we find out the format of a Mat object? So that i can choose the proper converting selection for it.

Vladislav, when i tried the way you described i am getting this error: Assertion failed (scn == 2 && depth == CV_8U).

Its normal short, so i changed it to CV_16SC3 , but this time it crashed with an error that says : " Assertion failed (depth == CV_8U || depth == CV_16U || depth == CV_32F)" in cvColor()

Also i can see the picture. When i use mt.data() and send it to the Java part as byte[] and create bitmap, i can see the picture in color.

This is how i get the pixels from vuforia:

short* input= (short*) imageRGB565->getPixels();
int width = imageRGB565->getWidth();
int height = imageRGB565->getHeight();

I changed it like that, but still it crashes at the same point.

Hi again Mathieu, the app crashes when i try to convert the RGB to GRAYSCALe like this:

Mat mt= Mat(width, height, CV_16UC3, pixels);
Mat graymat;
LOG("Before convert , image type %d",mt.type());
cvtColor(mt,graymat,CV_BGR2GRAY);

It crashes during the cvtColor method. And the LOG gives the image type as 18. Is it a normal value? What do you think the problem can be

Hi Mathieu, Thank you for the quick answer. Yeah the image is in color. I will try that.

Hello I am using Vuforia for creating an Augmented Reality Android app. And i want to do image processing on the images i get from Vuforia's camera frames. I am thinking of using OpenCV natively for that only purpose.

Vuforia captures images with its own Image class. but you can get pixels from that image as " short*". So I guess i have to turn those pixels somehow to Mat class so that i can use the image processing methods on it. But i couldnt find how to do that?

Can someone help me about it? Some suggestions

Thanks