I am attempting to find the center of several sets of 3D points on a sphere. Each set is comprised of three or more points that fall on the arc of a circle, but not perfectly as they have been supplied by an object detection algorithm, so there is some inherent error in these points. This is for me where the difficulty lies, I cannot simply solve the equations, I need to try and minimize variance in radius to this point across all three-point sets.
Currently, I am calculating a plane of best fit for each set of points. By calculating the radius (perpendicular distance) to this normal for each set and determining the variance I can figure out which plane (normal or center of rotation) fits all three sets the best. I am also doing this for an average of the three planes and for two planes after throwing out the plane that agrees least with the other two. So I am getting a pretty decent approximation currently.
My question is, does anyone know how to implement in Python some sort of function that can help me find a normal vector through these points that minimize the variance in radius for all sets. I suspect this won't be far off my current approximation, but am looking for the most accurate solution to this problem.
The picture below shows the results of what I am currently doing. The pink points represent the points I am using, labeled 0,1,2 for each set of points. The blue dots represent the normal vector projected to the surface of the sphere. The orange is the average of the three blue dots projected to the surface of the sphere. Ignore green they are not relevant to this. To minimize the variance my code is currently telling me that axis (blue dot) 0 results in the least variance in radius for the data set as a whole, but I highly doubt it is the best fitting point.
Related
I am trying to solve the following problem in Python. The problem comes from an image processing problem when i use the Finite Element Method.
In my problem, I have a set of triangles and a ray. Each triangle consists of three 3-D points, and the ray is in the form of a 3-D point and a 3-D vector. How can I determine the first triangle that is passed through by the ray? Now I do not even have an algorithm for this. Any inputs will be appreciated.
The first thing I would do, is translate the whole data set, subtracting the 3D ray origin. Then rotate the data set so that the ray's 3D vector aligns with the X-axis. See How to find the orthonormal transformation that will rotate a vector to the x axis?.
Now the problem has been converted to filter for triangles that cross the X-axis with a non-negative X-coordinate, and among those find the one whose crossing point has the minimal X-coordinate. So
For each triangle check where its plane crosses the X-axis. See Determine point of interesction of plane with axis given points of plane
Then throw away the triangles where that crossing point (on the X-axis) is not within the bounds of the triangle (check for each of the three edges that this point is at the "inner" side of it). See Check whether a point is within a 3D Triangle
Throw away the triangles whose crossing point has a negative X-coordinate.
Among the remaining triangles (that really cross the X-axis on the positive side) find the one with the minimum crossing point in terms of X-coordinate.
I have about 300,000 points defining my 3D surface. I would like to know if I dropped a infinitely stiff sheet onto my 3D surface, what the equation of that plane would be. I know I need to find the 3 points the sheet would rest on as that defines a plane, but I'm not sure how to find my 3 points out of the ~300,000. You can assume this 3D surface is very bumpy and that this sheet will most likely lie on 3 "hills".
Edit: Some more background knowledge. This is point cloud data for a scan of a 3D surface which is nearly flat. What I would like to know is how this object would rest if I flipped it over and put it on a completely flat surface. I realize that this surface may be able to rest on the table in various different ways depending on the density and thickness of the object but you can assume the number of ways is finite and I would like to know all of the different ways just in case.
Edit: After looking at some point cloud libraries I'm thinking of doing something like computing the curvature using a kd tree (using SciPy) and only looking at regions that have a negative curvature and then there should be 3+ regions with negative curvature so some combinatorics + iterations should give the correct 3 points for the plane(s).
I have a numpy array depicting a one-pixel wide, discrete, connected curve. This curve is obtained by the Skeletonization operation of image processing. I am trying to find the curvature of the above curve at an arbitrary point, to detect bends/kinks (which will have high curvature value).
I tried to implement the above using the general formula for curvature. However, since this a pixelated, discrete curve, whose generating function is unknown, I tried to resort to using numpy gradient instead.
The problem I see with the above is that, since the curve is one-pixel wide, at any point the slope can be only one of 0, 1 or infinity. As a result, the curvature values that I get are mostly meaningless or useless.
I am looking for some suggestion on where to start in order to get a smooth curve out of the above, so that I can calculate curvature in a more meaningful way. Can somebody suggest any mathematical operation or convolution that I can apply to achieve the same? Below is a representative binary image that I have.
P.S. I am very, very new to image processing, so references to standard algorithms (in math books) or library implementations would be very helpful.
An established way to do this is to fit a low-order parametric curve to each of the skeletonized points using two or more neighbouring points. Then you compute curvature at the point using the fitted curve parameters with an analytic formula. Several curve models can be used. The two main models are:
A circle. The radius of curvature, R is the reciprocal of the curvature. For a curve, it equals the radius of the circular arc which best approximates the curve at that point. You can fit a circle to a set of 2D data points using various methods. A python library that has implemented several is here.
A quadratic. This can be fitted to the point and its neighbours, then curvature can be estimated through second-order differentiation of the curve here. You can use numpy.polyfit to fit this model. A simple strategy is to first estimate the tangent vector at the point, by fitting a local line (e.g. with polyfit using an order 1 curve). The you rotate the points to align the tangent vector with the x axis. Finally you fit a 1D quadratic f(x) to the rotated points using polyfit.
The tricky thing with making any curvature estimator is that curvature can be estimated at different scales. For example, do I want my estimator to be sensitive to high frequency detail or is this actually noise? This decision manifests in the choice of neighbourhood size. Too small, and errors from noise and discretization lead to unstable estimates. However too large, and there may be large modelling error (error by approximating the curve as a parametric function). Generally you have to select the best neighbourhood size yourself.
You're also going to have some poor curvature estimates at junction points, but that's largely unavoidable as curvature is not well defined there. A naïve fix could be to segment all paths at junction points, and then estimate curvature on each path individually.
Toby gave an excellent suggestion regarding junction points: detect the junction points and take each line in between those independently.
Detecting junction points (and end points). This is quite simple: all pixels that are set and have more than two neighbors are junction points. All pixels that are set and have exactly one neighbor are end points. Detect all those points and put their coordinates in a list.
Finding the lines in between pairs of points. Starting at each coordinate in your list, look for a line starting there. Note that for the junction points, you'll have at least three lines starting there. If you do this, you'll find each line two times. You can remove duplicates by reversing the lines that end to the left of where they start (and if the two end points are on the same image column, take the one on top as the start). Now they will be directly comparable, so you can delete the duplicates (or not store them in the first place). Note that just comparing start and end point is not sufficient as you can have different lines with the same start and end points.
Tracing each line. The step above requires that you trace each line. See if you can figure it out, it's fun! Here is a description of an algorithm that traces the outline of objects, you can use it as inspiration as this problem is very similar. Store a vector with x-coordinates and one with y-coordinates for each line.
Smoothing the lines. As you noticed, consecutive steps are in one of 8 directions, so angles are strongly discretized. You can prevent this by smoothing the coordinate vectors. This is a quick-and-dirty trick, but it works. Think of these vectors as 1D images, and apply a smoothing filter (I prefer the Gaussian filter for many reasons). Here you filter the vector with x-coordinates separately from the vector with y-coordinates.
Computing the curvature. Finally, you can compute the curvature of the curve, as the norm of the derivative of the unit normal to the curve. Don't forget to take the distance between points into account when computing derivatives!
I have boundaries of semi-circle or ellipse shaped objects. Example image is
The boundary can be slightly jagged (when you zoom in). I am looking to detect a point of interest (location x and y) on these curves, where we see a definite change in the shape, such as
There can be two outputs:
No point of interest: we cannot find specific features
Point of interest with x and y location
Currently, I am using Python and OpenCV. I cannot think of a efficient and effective way to solve this problem. Any help will be really appreciated.
Nothing says that others will agree with my closure vote, so ...
I suggest two steps:
Fit an ellipse to the given points. I'm sure you've already found curve-fitting algorithms (and perhaps software packages) by now -- and asking for those is specifically proscribed on Stack Overflow.
Code a small anomaly detector, which works on the difference between the fitted curve and the actual data points.
Step 2 depends heavily on your definition of "point of interest". What are the criteria? I notice that your second point of interest actually lies very close to the fitted curve; it's the region on either side the deviates inward.
I suggest that you do your fitting in polar coordinates, and then consider the result in terms of theta and radius. Think of "flattening" the two curves as a single unit, so that the central angle (theta) is the new x-coordinate, and the distance from the center is the new y-coordinate.
Now, subtract the two curves and plot the difference (or just store this new curve as an array of points). Look for appropriate anomalies in these differences. This is where you have to decide what you need. Perhaps a sufficient deviation in the "r" value (radius, distance from center"); perhaps a change in the gradient (find a peak/valley, but not a gently sloping bulge). Do you want absolute difference, or an integral of deviation (area between the fit and the anomaly). Do you want it linear or squared ... or some other function? Does the width of the anomaly figure into your criteria?
That's what you need to decide. Does this get you moving?
I have a set of points extracted from an image. I need to join these points to from a smooth curve. After drawing the curve on the image, I need to find the tangent to the curve and represent it on the image. I looked at cv2.approxPolyDP but it already requires a curve??
You can build polyline, if order of points is defined. Then it is possible to simplify this polyline with Douglas-Peucker algorithm (if number of points is too large). Then you can construct some kind of spline interpolation to create smooth curve.
If your question is related to the points being extracted in random order, the tool you need is probably the so called 2D alpha-shape. It is a generalization of the convex hull and will let you trace the "outline" of your set of points, and from there perform interpolation.