I want to use a camera as a Document Scanner and I need to fix the distortion.
I have seen solutions that allow the user to point on the corners of the document, and they fix the distortion according the points.
I am trying to do that with OpenCV, but I cannot find a simple way to do that.
Can somebody show me how o do that with simple 4 corners of the document?
Thank's
Meir
I found the solution here: nuigroup.com/forums/viewthread/3414 that's exactly what I was looking for.
Related
I had a similar situation like described in this question. Within the question, the user who asked wrote the following equation:
It seems to work *, but does somebody know, how to get this equation? I was not able to find anything like this anywhere else...
Thanks in advance!
*) I also guess I kinda understand the equation:
the left part is the ray coming from the camera and going through the image point, therefore on the right side the real-world coordinates have to be translated and rotated as if the camera would be placed in (0,0,0)... sort of?
I just can't find a paper or anything to confirm the equation
It is a simple pinhole camera model equation. I guess what you are looking for, at least partially described PDF: here. You can also google more math yourself using keywords: camera calibration, pinhole model.
I want to find out a given image in an input image......first I used template matching using opencv but its not giving proper result. So I switched to keypoint matching using SIFT with opencv. But I am not able to find pixel location in a input image. Someone please help me I am looking everywhere but not able to solve my problem
It is a multifaceted problem. Its easier to help with specific questions or problems.
This tutorial on Youtube helped me with this problem.
I'm trying to detect an object composed of other objects. Actualy, there are three circles in my binary image which shape up a triangle as shown here:
These circles are correctly detected, but only as single objects as shown here:
What I need to have is an aggregation or composition of these objects, so they get detected as one big object as shown here:
The bigger goal is to get the image moments to get the rotation and scale of the shape. Please share your ideas or code if you have any, it would be well appreciated.
I would suggest using the bounding box functions of opencv
Here is a link to an example of bounding box in C++ OpenCV, however if you are using something like Python, it might be worth your while looking at this link, which is a full set of tutorials for working with binary images and contours (including bounding box/elipse)
Again if you are using the Python port, look at this set of tutorials, they really are great and have a massive supply of information on most functions of OpenCV.
Hope this helps.
Good luck.
Your question is very similar to this question, which has answers with code examples. Alternatively check the documentation of OpenCV. If you are interested in the convex hull of your points, see cv::convexHull().
I'm having the user capture an image of a barcode that they are asked to place inside a bounding box drawn on the camera overlay. I'd like to detect whether that barcode was properly placed in the box. What would be the best way to detect this? OpenCV's houghLines? And how would I figure out if the lines detected from houghLines are the barcode's corner and that they match up with the camera overlay's bounding box?
Thanks in advance for any help
The post How to find corners on a Image using OpenCv has the solution for your problem. I did the same for my project and it worked like a charm :) If need help in understanding or tuning the algorithm, don't hesitate.
Cheers
I want to achieve coloring picture effect on the photo and asking for help with computer vision techniques as I am newbie in this area. I started out with bilaterial filter and min filter. But I am stucked what to do next. I use opencv with python for prototyping. Is there any patterns that might help me to achieve this? Appreciate your help and time.
One of the possible solutions:
Convert image to grayscale
Apply Canny edge detector or Sobel filter
Hope it helps.
What you are looking for is called "Non-photorealistic rendering". In its simplest form, it is what Astor suggested you. But there are many studies in this field, starting with automatic creation of sketches to complex cartoon-like effects and more.
Here is a paper on this topic, and another one here, but you can find more by googling "Non photorealistic rendering"