I want to draw a "bounding box" around a skewed rectangle. I thought I could use the cvMinAreaRect2() function but it only handles the rotation, see this image:
Is there any function to solve this?
If not, any ideas how to implement it?
Compute both MinAreaRect and ConvexHull. Then, for each of the four points found by MinAreaRect, find the corresponding nearest point in the convex hull.
Related
I have a couple of spherical images, given in equirectangular projection, looking at the same object from different positions. I know the absolute pose of each image e.g. position in geographical coordinates and roll/pitch/yaw angles. Given the pixel coordinate of a point in one image I would like to find a way to draw the epipolar line (where the correspondent point lies) in the other one.
I tried to deal with Essential/Fundamental matrix in python using OpenCV but I did'nt figure out how to achieve this.
Any help is really appreciated.
Thanks
There a is an ellipse on the picture,just as following.
I have got the points of the contour by using opencv. But you can see the pictrue,because the resolution is low, there is a straight line on the contour.How can i fit it into curve like the blue line?
One Of the method to solve your problem is to vectorize your shape (moving from simple intensity space to vectors space).
I am not aware of the state-of-art in this field. However, from school information, I can suggest this solution.
Bezier curves, you can try to model your shape using simple bezier curve.This is not a hard operation you can google for dozen of them. Then, you can resizing it as much as you want after that you may render it to simple image.
Be aware that you may also Splines instead of Bezier.
Another method would be more simple but less efficient. Since you mentioned OpenCV, you can apply the cv::fitEllipse on the points. Be aware that this will return a RotatedRect which contains the ellipse. You can infer your ellipse simply like this:
Center = Center of RotatedRect.
Longest Radius = The Line which pass from the center and intersect with the two small sides of the RotatedRect.
Smallest Radius = The Line which pass from the center and intersect with the two long sides of the RotatedRect.
After you got your Ellipse Parameters, You can resize it as you want then just repaint it in the size you want using cv::ellipse.
I know that this is a pseudo answer. However, I think every thing is easy to apply. If you faced any problem implementing it, just give me a comment.
I'm converting a KML file to a polygon stored on my database thanks to RGeo
I need the centroid: #area.centroid and I would like to find the approximative radius on my shape.
I don't know if it's possible. My database field is a geometry. That means I can store polygons AND multi-polygons. The type of my area is RGeo::Geos::CAPIMultiPolygonImpl
Finally I think I need the check if the area is a simple polygon or a multiple. If it's a simple one, I would like to find the radius. Else nothing.
Thank you for your help !
I finally did like below:
Convexe Hull
Centroid of this Convexe Hull
minimum distance between Convexe Hull points (thanks to exterior_ring function) and the centroid
By the way I don't know why I had to take the minimum. It was to adapt my scale. I don't really understand how the centroid of a polygon (which could have really strange shape) can be found.
Anyway, it's working thanks to the 3 steps written previously.
how do you get the centroid of an irregular shape using OpenCV?
I'd recommend looking at the cv::Moments (C++) or cvMoments (C) function.
This StackOverflow thread gives some example code for a problem very similar to yours.
This post goes into some of the theory related to finding object center-points.
What do you mean by centroid?
If it's the center of mass, you can compute the average of the coordinates of the points that are inside your shape. But the center of mass can be outside the shape, for "irregular" (non-convex) shapes.
If you want the point inside the shape that is the further away to any of the contour point, you can have a look at distTransform function.
I have to detect the pattern of 6 circles using opencv. I have detected the circles and their centroids by using thresholding and contour function in opencv.
Now I have to define the relation between these circles in a way that should be invariant to scale and rotation. With this I would be able to detect this pattern in various views. I have to use this pattern for determining the object pose.
How can I achieve scale/rotation invariance? Do you have any reference I could read about it?
To make your pattern invariant toward rotation & scale, you have to normalize the direction and the scale when detecting your pattern. Here is a simple algorithm to achieve this
detect centers and circle size (you say you have already achieved this - good!)
compute the average center using a simple mean. Express all the centers from this mean
find the farthest center using a simple norm (euclidian is good enough)
scale the center position and the circle sizes so that this maximum distance is 1.0
rotate the centers so that coordinates of the farthest one is (1.0, 0)
you're done. You are now the proud owner of a scale/rotation invariant pattern detector!! Congratulations!
Now you can find patterns, transform them as suggested, and compare center position & circle sizes.
It is not entirely clear to me if you need to find the rotation, or merely get rid of it, or detect if the circles actually form the pattern you linked. Either way, the answer is much the same.
I would start by finding the two circles that have only one neighbour. For each circle centroid calculate the distance to the closest two neighbours. If the distances differ in more than say 10%, the centroid belongs to an "end" circle (one of the top ones in your link).
Now that you have found the two end circles, rotate them so that they are horizontal to each other. If the other centroids are now above them, rotate another 180 degrees so that the pattern ends up in the orientation you want.
Now you can calculate the scaling from the average inter-centroid distance.
Hope that helps.
Your question sounds exactly like what the SURF algorithm does. It finds groups of interest and groups them together in a way invarant to rotation and scale, and can find the same object in other pictures.
Just search for OpenCV and SURF.