I am learning point cloud data processing and currently trying to do cylinder fitting.
My aim is to take the point cloud, pick a subset of points and on them do least squares fitting of the cylinder. However, unlike a plane, I was not able to find a way to fit it.
Is there some straightforward way to do it? Or is there some better algorithm for the job than least squares?
Related
I'm trying to detect squares using Point cloud library. I have pcl data from a 3D lidar in which I need to find squares. Ransac doesn't have a model for square. I wish to know what can be the most efficient method for square detection.
If you are looking for a filled square, the SACMODEL_PLANE should be able to find it. You may need to cluster the inliers of the plane model, and filter the clusters to find the location of the square.
If you are looking for the outline of a square, the SACMODEL_LINE should be able to find the 4 sides separately. You will then need some logic to filter out lines that do not belong, as well as to combine the inliners of the correct lines.
I have a list of (x, y) points. I know how to make a list of Bézier curves which pass through all of those points and have a continuous first (and second, though less important) derivative. However, the list that I end up with is far too long. I would prefer to approximate the points I have if it lets me cut down on the number of curves I have. I would like to be able to pass a parameter of either how close an approximation I get or a maximum number of curves, preferably the former.
The reason I want this is that the end result will have a graphical UI where users can edit the Bézier curves, and it isn't critical that the curves pass through each point exactly, as long as they are close. More curves makes it harder to edit.
EDIT:
Some more information about the purpose of this. I'm trying to make image editing software. When someone loads a bitmap, I want to be able to trace a center line. Potrace is what I would use to trace the outline of a shape, but it won't work for tracing strokes. I've been able to identify lots of points along the center line, and I want to turn this data into a list of connected Bézier curves. The reason I don't want to make a Bézier spline is that there will be too many control points for this to be easy to edit. "Too many" is not an easy-to-define term, but I would like to be able to pass a parameter to limit the number of curves. Either a function that minimizes how far the curves are from the points based on a maximum number of curves or a function that minimizes the number of curves based on a maximum deviation from the points.
Several approaches exist for achieving what you want to do:
1) Use RDP algorithm to reduce the number of points, then create a list of Bezier curves passing thru the remaining points.
2) Use curve fitting algorithms (for example, Schneider algorithm) to produce multiple Bezier curves that are connected with G1 (tangent) continuity. Check out Schneider algorithm implementation in this link.
3) Use least square fitting with B-spline to produce a single B-spline curve.
From implementation point of view, approach 1 is probably the easiest one for you as you already know how to create Bezier curves interpolating a list of points. Approach 3 will be much more difficult to implement and you probably will have to convert the B-spline curve into Bezier curves so as to use them at the UI level. Please refer to this SO article for detailed discussion.
I am currently working on a experiment that i took multiple photos
of a scene on diferent days with a fixed camera position.
The problem is that on real world it is hard to keep the camera
perfectly fixed.
What i need is to fix the small variance I got automaticaly. The research
I made returned methods considering more complex assumption, like camera
pose estimation, homography estimation etc.
For me its enought to discover just the movement at the image plane returning an
x and y.
A perfect solution would be a function such as:
function [movx movy] = detectMotion(im1,im2).
The solution I already made was to calculate some image features, like harris or
hessian, match them and after manualy select the best ones and use the difference
of their position as a camera displacement estimation. I dont know if this is good
enough but it would be better if it was made automaticaly.
You can do the feature matching automatically be extracting feature descriptors around the interest points. Take a look at this OpenCV tutorial on how to perform feature matching using SURF and FLANN. Once you have the feature matches, run RANSAC or least squares to find the best fit for an x- and y-offset. This will give you a decent estimate of the camera motion.
Another option is to compute sparse optical flow on the detected interest points between the two frames, followed by the RANSAC or least squares procedure as above to compute the best x- and y-offset. Dense optical flow could possibly be more accurate, but at the same time could prove to be overkill.
Mapping a point cloud onto a 3D "fabric" then flattening.
So I have a scientific dataset consisting of a point cloud in 3D, this point cloud comprises points on a surface that is curved. In order to perform quantitative analysis I however need to map these point clouds onto a surface I can then flatten. I thought about using mapping tools sort of like in the case of the 3d world being flattened onto a map, but not sure how to even begin as I have no experience in cartography and maybe I'm trying to solve an easy problem with the wrong tools.
Just to briefly describe the dataset: imagine entirely transparent curtains on the window with small dots on them, if I could use that dot pattern to fit the material the dots are on I could then "straighten" it and do meaningful analysis on the spread of the dots. I'm guessing the procedure would be to first manually fit the "sheet" onto the point cloud data by using contours or something along those lines then flattening the sheet thus putting the points into a 2d array. Ultimately I'll probably also reduce that into a 1D but I assume I need the intermediate 2D step as the length of the 2nd dimension is variable (i.e. one end of the sheet is shorter than the other but still corresponds to the same position in terms of contours) I'm using Matlab and Amira though I'm always happy to learn new tools!
Any advice or hints how to approach are much appreciated!
You can use a space filling curve to reduce the 3d complexity to a 1d complexity. I use a hilbert curve to index lat-lng pairs on a 2d map. You can do the same with a 3d space but it's easier to start with a simple curve for example a z morton order curve. Space filling curves are often used in mapping applications. A space filling curve also adds some proximity information and a new sort order to the 3d points.
You can try to build a surface that approximates your dataset, then unfold the surface with the points you want. Solid3dtech.com has the tool to unfold the surfaces with the curves or points.
This is something related with Mathematics as well. But this is useful in computing as well.
Lets say you have 10 coordinates. (x1,y1)(x2,y2)..... in 2D Space. (i.e on a X-Y Plane). Can we find a single smooth curve going across the each coordinate.
While expanding the question, If the space is 3D, then can we find an equation of a smooth surface that going across a given set of spacial coordinates?
Are there any Libraries (Any language) \ tools to perform such calculations?
What you should be looking for is some library implementing NURBS (or Non Uniform Rational B-Splines). This will solve your problem in both 2d and 3d, since 2d is just a special case of the 3d.
Roughly speaking, you are not interested in the actual equation, you are only interested in getting the points approximated with smooth curves or surfaces. This is done by finding "control points" in 2d or 3d space, which are multiplied with B-spline base functions. A NURBS library will do this for you.
Cheers !
Edit:
Have a look at this one
you can always fit an order-10 polynomial through the points. that's not necessarily what you want to do, though - fitting a smooth curve via a series of splines will give you a better-looking result. the curve-fitting article on wikipedia gives you a good overview of the various options.
In the 2D case you are asking for curve fitting. This actually exists in excel, where you plot your points (I usually use XY scatter if you have x and y listed) and then right-click on the curve. Select Add Trendline. There you can choose which kind of function you want to fit to and you can ask excel to display it in the image (Tab named Options, check the box "Display equation on chart"). Nice and quick.
Otherwise you can use matlab and use the lsqr (least square method). If you want to find the polynomial closest that best describes your data you could use the polyfit function. It uses the least square method, but returns coefficients. Matlab has a whole set of other algorithms for solving/finding "best" approximations to systems of linear equations. I mention lsqr because it is one of the simplest to implement yourself if you don't have matlab. On the other hand it is for solving sets of linear equations - I don't know anything about your data.
Have a look at splines
in wiki
an interactive introduction
Searching for 'spline interpolation library' might give some useful hints for implementations.