Soarez's profile - activity

Hi i got it to work for me, u can try the same =). Since ther is no function to compute pose on fisheye lens, and the so

Hi, you manage to get it working? I had the some problem .

Can you see the image? probably you have a monochromatic image in gray scale. Dim = 2, means that you dont have a x,y m

Remap a single point unsing maps from initUndistortRectifyMap(). Hi, im tring to remap a set of points. Im using fisheye

Remap a single point unsing maps from initUndistortRectifyMap(). Hi, im tring to remap a set of points. Im using fisheye

Remap a single point unsing maps from initUndistortRectifyMap(). Hi, im tring to remap a set of points. Im using fisheye

Hi, i need to do the same. Did you get any sucess to remap a single point or an array of points?

thank you , i go check it out!

Hello, i wanna apply a Gaussian kernel to the image, but the thing is .... just to a certain part. I have a big image, 2064/1544 +/-, most of the image is black and i don't need it. Gaussian blur apply th kernel to the full image and takes time, although i have made a custom Gaussian for my problem.

My cycling to each pixel : 1. check if pixel its over a Threshold value. 2. if positive apply Gaussian kernel to the surroundings.

My code works correctly and if i get a TH value to 0, i got the same result as the Gaussian blur function, Gaussian Blur is always faster.

Question: there is already a function like this!? Question: If not , how can i optimize my code!?

The A) worked

fish eye methods give me this http://answers.opencv.org/question/13...

Im using a +/-86º lens and running the calibration demo, provided by openCV installation , and the undistorted images look like this:

some one knows why?

Im trying to calibrate a camera using openCV example, and the calibration finish without error but when is showing the undistorted images doesn't look good. im getting images like!

its a 2048 by 1540. and the lens has +/-86º of opening.

/*******SOLVED***********/

Used normal calibration (not fisheye) Take more 20 photos with aggressive angles. Callibrated with all the images and got this result, good for now =) .

@berak , its working , used the same think that i used on the question , because the images is 8bit packed RGB, so im sorry for your time , but thanks again =)

OK, i think that i have a separated planar image like you said on your answer. But what if i have a mixed memory where the 3 first address are for pixel one, the next 3 for pixel 2, the next 3 for pixel 3 and so one. see on the link http://imgur.com/a/19Ma2 , red green and blue order are just for the example.http://imgur.com/a/19Ma2 (img)

yes i printed it, padding values are 0 , so no problem =).

just one more thing @berak , and already thanks a lot for your help, if i have a image that come in like address 1 - pixel 1 blue address 2 - pixel 1 reg address 3 - pixel 1 green address 4 - pixel 2 blue address 5 - pixel 2 reg address 6 - pixel 2 green and so on , how do i initialise the mat!?

yeah, i already asked the seller waiting for response. Do you know what padding bytes mean!? for each image i recive a x_padding and y_padding, but i dont know what they mean.

im getting images like this ...link text , do you have any idea why!?
http://imgur.com/a/8bXEG

@berak so i cant do it on the thread?

@berak , Im working on a thread, and on the main i can load images but on the thread i can't, and the same with display and imwrite.

im just doing

namedWindow( "Display windsssow", WINDOW_AUTOSIZE );// Create a window for display.
imshow( "Display windsssow",bgr);
cv::waitKey(3);

it should work =)

Looks like i have a 3 channel image now ,i print one col ...

although i cant see the image on imshow, but this could be other problem.

Hi, Im working with a camera, and the camera gives me the image in its own type of variable, that contains image width , height , depth , and other things.

In this data i receive a pointer for the first pixel , and i know that the RGB is organized like (assuming the image as n pixels), the first n address have the R, the next n address have the G and the last n address have the B.

So im trying to initialize the image like this.

cv::Mat dummy_query = cv::Mat(img->h, img->w, CV_8UC3, pData , 0 );

pData is the pointer from the first pixel. the 0 ,on openCV doc they say its the step , i really dont understand what is this step.

whit this i get some data to the matrix, but when i do the imshow nothing happens.

hope someone can help. =)