Point Cloud Library (PCL) Developers mailing list This forum is an archive for the mailing list pcl-developers@pointclouds.org ( more options ) Messages posted here will be sent to this mailing list. Pcl:: transformPointCloud (const pcl:: PointCloud & cloud_in, pcl:: PointCloud & cloud_out, const Eigen:: Affine3f & transform) If we are using PointT = pcl::PointXYZ, both of these methods have the desired effect, however, if we use PointXYZI, or XYZRGB, only the second method will copy the I or RGB data into the second cloud.

• Point Cloud Library Pcl
• Point Cloud Classification With Pcl
• Point Cloud Library Pcl Users Mailing List

>-Original Information- >From: Justin Rosen mailto: >Sent: Monday, Sept 26, 2011 6:21 PM >To: Julius Kammerl >Subject matter: Re also: PCL-users Making use of the octree classes >>Thanks a lot Julius! I believed this had been the situation. I experienced to learn up on my >bitwise math to discover that this produced sense. Simply needed to create sure as >a lot of this stuff I'm returning to from my university times from 5-6 decades >back. In inclusion, the ray casting worked well which could only suggest that >it was in reality in the exact same purchase. >>Thanks a lot for getting so individual with me and clarifying things associated to >all my questions! I enjoy the help:) >>As an aside, I don't think I'll be needing all that mumbo jumbo about >non pillow bounding containers octrees.

Probably it's quicker in some situations? >But for my requirements, your things just functions! In addition, I finally emerged >to the conclusion of one of your prior comments: >>'By description, every part node of the octree describes a bounding container >of underlying nodes and the basic node encapsulates all points.' >>I study this many situations and had been like, yeah, but the divisions still >don't shop the beliefs. And after that it hit me, you have got to subdivide >on the way lower! >>Thanks a lot Again, >Justin >>>>On Mon, Sep 26, 2011 at 2:20 AM, Julius Kammerl >published: >>Hi Justin, >>>>Your number of the octree kid order appears to be right.

>>>>This is definitely how the child index is certainly calculated: >>>>childIdx = ((!!(keyarg.a depthMaskarg)) >((!!(kéyarg.y depthMaskarg)) >(!!(kéyarg.z . depthMaskarg)); >>>>depthMaskarg is a bitmask including a one bit which is usually altered >>matching to the depth degree in the octree. >>>>Cheers, >>Julius >>>>>>>>>-Original Information- >>>From: Justin mailto: >>>Sent: Weekend, September 25, 2011 5:52 Was >>>To:; Point Fog up Library (PCL) Users >>>Subject: Re also: PCL-users Making use of the octree classes >>>>>>Will anyone understand in which purchase child nodes of a department are stored >>>within 0ctreeBase? >>>>>>PCL-developers maiIing list.

Hey JuIius, Thanks! I wear't need this at the second, but will end up being helpful as I start providing more visualization feed-back. The only issue I see with the node iterator is usually that department nodes do not store their bounding boxes. As you've mentioned >>'By description, every department node of the octree defines a bounding box >>of underlying nodes and the main node encapsulates all factors.' This is usually true when you itérate through each nodé and children subdividing as you move. This seems harder to deal with with an iterator zero? Lets say I'meters a pcl-usér who doésn't possess gain access to to the internaIs of the class and I put on't strategy on creating my own.

How could I attract or imagine the octree at each level? Probably we can'capital t and this execution requires to be composed subclassing the present foundation octree? On Sép 27, 2011, at 6:43 Was, Julius Kammerl published.

Forwarded information - From: engin Time: Fri, Mar 17, 2017 at 8:55 AM Subject matter: PCL-users pcl::PCA eigenvectors/eigenvalues are computed with incorrect Cov matrix? (1.8) To: Hello, Collection 85-88 of: /./ Eigen::Matrix3f alpha dog = staticcast (clouddemean.topRows. clouddemean.topRows.transpose ); // Compute eigen vectors and ideals Eigen::SelfAdjointEigenSolver evd (alpha); /./ leader is most probably the covariance matrix as it is certainly approved to SelfAdjonitEigenSolver. But shouldn't that end up being: leader = 1/(N-1). clouddemean.

clouddemean.transpose? Not really placing the 1/(In-1) makes the eigenvalues reliant on the quantity of factors you have in the cloud. Double the amount of data points you have got in the specific exact same x,y,z period of time you will obtain eigenvalues that are twice simply because huge. The eigenvectors may nevertheless be good as they are usually furthermore normalized. But it still looks incorrect to me. Right here is usually an illustration, I replicated from: If cov is determined as in the program code, the eigenvalues you obtain will become: 0.44354, 80.79646' their percentage: M1/L2 = 182.16 The right values are: 0.049083, 1.284028 T1/L2 = 26.160 Was generally there a reason to create it like this that I'm lacking?

Engin - Watch this message in circumstance: Sent from the Point Fog up Library (PCL) Customers mailing list mailing list archive at Nabble.com. / PCL-developers mailing list.

>Forwarding this to the developers list because it appears like a serious >issue and it hasn't picked up any responses on the users list in a month. >>Any ideas? >>David >>- Forwarded message - >From: engin >Day: Fri, Mar 17, 2017 at 8:55 AM >Subject matter: PCL-users pcl::PCA eigenvectors/eigenvalues are calculated with >incorrect Cov matrix?

(1.8) >To: >>>Hi, >>Line 85-88 of: >>typical/impl/pca.hpp >/./ >Eigen::Matrix3y alpha dog = staticcast >(clouddemean.topRows. clouddemean.topRows.transpose ); >>// Compute eigen vectors and values >Eigen::SelfAdjointEigenSolver evd (alpha); >/./ >>alpha is most probably the covariance matrix as it will be handed to >SelfAdjonitEigenSolver.

But shouldn't that become: >>alpha dog = 1/(N-1). clouddemean. clouddemean.transpose? >>Not placing the 1/(In-1) can make the eigenvalues dependent on the number of >factors you have got in the cloud.

Twice the quantity of information points you possess in >the precise same x,y,z interval you will obtain eigenvalues that are double as >huge. The eigenvectors may still be great as they are also normalized. But it >still looks wrong to me.

>>Right here will be an example, I duplicated from >>principalcomponents.pdf>>: >>If cov is usually computed as in the program code, the eigenvalues you get will be: >0.44354, 80.79646' >their percentage: M1/L2 = 182.16 >>The appropriate values are usually: >0.049083, 1.284028 >L1/L2 = 26.160 >>Had been now there a cause to create it like this that I'michael missing? >>Engin >>>>- >View this information in framework: >PCA-eigenvectors-eigenvalues-are-computed-with-incorrect- >Cov-matrix-1-8-tp4044153.html >Sent from the Point Fog up Library (PCL) Customers mailing list mailing list >save at Nabble.com. >>/ >>>PCL-developers mailing list >>>PCL-developers mailing list. >Forwarding this to the programmers list because it seems like a critical >problem and it hasn't received any response on the users list in a 30 days. >>Any ideas? >>James >>- Forwarded message - >From: engin >Time: Fri, Mar 17, 2017 at 8:55 Are >Subject: PCL-users pcl::PCA eigenvectors/eigenvalues are usually computed with >incorrect Cov matrix? (1.8) >To: >>>Hi there, >>Range 85-88 of: >>typical/impl/pca.hpp >/./ >Eigen::Matrix3y alpha = staticcast >(clouddemean.topRows.

clouddemean.topRows.transpose ); >>// Compute eigen vectors and ideals >Eigen::SelfAdjointEigenSolver evd (leader); >/./ >>leader is presumably the covariance matrix as it is definitely passed to >SelfAdjonitEigenSolver. But shouldn'capital t that end up being: >>alpha dog = 1/(N-1). clouddemean. clouddemean.transpose? >>Not really placing the 1/(N-1) makes the eigenvalues reliant on the amount of >factors you have in the cloud. Double the quantity of data factors you have in >the precise exact same x,y,z time period you will obtain eigenvalues that are double as >big. The eigenvectors may still be fine as they are usually furthermore normalized.

But it >still looks wrong to me. >>Here is certainly an instance, I replicated from >>principalcomponents.pdf>>: >>If cov will be determined as in the code, the eigenvalues you obtain will become: >0.44354, 80.79646' >their ratio: M1/L2 = 182.16 >>The right values are usually: >0.049083, 1.284028 >L1/L2 = 26.160 >>Had been now there a cause to compose it like this that I'm missing? >>Engin >>>>- >View this information in framework: >PCA-eigenvectors-eigenvalues- are-computed-with-incorrect- >Cov-matrix-1-8-tp4044153.html >Sent from the Point Cloud Library (PCL) Customers mailing list mailing list >save at Nabble.com. >>/ >>>PCL-developers mailing list >>>PCL-developers mailing list PCL-developers mailing list.

Offers anyone attempted the Point Cloud Compression guide?: I'michael trying it óut, but mine fréezes at PointCloudEncoder->encodePointCIoud (cloud, compressedData); Nó errors or anything, it simply stays there tugging on my Central processing unit and certainly not surface finishes this command word (I've waited 10 mins). Offers anyone else encountered similar issues?

The code I'michael using can be precisely the same as the tutorial except one point.I had to modify #consist of to #include Thanks a lot -Aaron. Hi there Aaron, Sorry to hear that you are usually experiencing problems with the pcl point cloud compression.

Which pcl edition are usually you using? I recommend to use the most recent trunk edition. Will opennistreamcompression (in apps) function on your machine?

It would become great to understand where exactly in the program code it freezes. Could you test to identify this execution loop? Which working system are usually you using (32bit/64bit)? I simply tested the short training program code with the most recent pcl trunk area and it appears to become working great on my machine.

Cheers, Julius -First Information- From: mailto: On Behalf Of airuno2l Put: Fri, June 17, 2011 9:28 PM To: Subject matter: PCL-users Point Cloud Compression guide Has anyone attempted the Point Cloud Compression tutorial?: ssion I'meters attempting it óut, but mine fréezes at PointCloudEncoder->encodePointCIoud (cloud, compressedData); Nó errors or anything, it simply sticks there pulling on my Processor and certainly not surface finishes this command word (I've waited 10 minutes). Provides anyone else encountered similar troubles? The code I'michael using is certainly exactly the exact same as the guide except one issue.I acquired to alter #include to #consist of Thanks -Aaron - View this information in circumstance: -Compression-tutoriaI-tp30649.html Put from the Point Fog up Library (PCL) maiIing list mailing Iist save at Nabble.com. / /. >Sorry for not really enough fine detail. I'michael making use of pcl 1.0 built from binaries on >windows vista-64bit making use of msvc express.

>>I will consider opennistreamcompression on Monday (I'meters home for the weekend break). >>The program code freezes at this command word. >>PointCloudEncoder->encodePointCloud (cloud, compressedData); >>Is certainly there a opportunity the guide only functions with the trunk area version?

I glanced >at the version of octreepointcloudcompression.h in both thé binaries ánd >in the trunk area and it seems to have changed a bit. >>- >Watch this information in circumstance: >Sent from the Point Fog up Library (PCL) maiIing list mailing Iist save at Nabble.com. >>/ >/. The stage compression guide should function with older versions, as well. I perform not know about any general problem here.

Point Cloud Library Pcl

Please allow me understand as shortly as you have some more information about this insect, therefore that we can repair it quickly! Regards, Julius -Original Information- From: mailto: On Account Of Radu T.

Rusu Sent: Weekend, Summer 19, 2011 12:27 Was To: Point Cloud Library (PCL) mailing list Subject: Re also: PCL-users Point Cloud Compression guide I'll allow Julius respond to the question in details, but in common, I would constantly try trunk for any fresh feature that hasn't been recently tested more than enough or provides been recently released, as any bugs/issues that the first edition might have acquired could have got happen to be be set. Cheers, Radu. Point Cloud Library (PCL) - On 02:14 Evening, airuno2m had written. >Sorry for not enough details. I'meters using pcl 1.0 constructed from binaries on >windows vista-64bit making use of msvc express. >>I will test opennistreamcompression on Monday (I'michael home for the weekend break).

Point Cloud Classification With Pcl

>>The code freezes at this order. >>PointCloudEncoder->encodePointCloud (cloud, compressedData); >>Can be there a opportunity the guide only works with the trunk area edition? I glanced >at the edition of octreepointcloudcompression.l in both thé binaries ánd >in the trunk and it seems to have transformed a little bit. >>- >See this information in circumstance: -Compression-tutoriaI-tp30042.html >Sent from the Point Cloud Library (PCL) maiIing list mailing Iist save at Nabble.com. >>/ >/ /. Hi, You'll find opennistreamcompression in the latest trunk version (in apps). You could furthermore test pclstreamcompression in creation/tools.

Point Cloud Library Pcl Users Mailing List

Cheers, Julius -Original Message- From: mailto: On Account Of airuno2d Sent: Mon, August 20, 2011 4:44 Evening To: Subject matter: Re also: PCL-users Point Fog up Compression short training Julius, I looked for opennistreamcompression but can't discover it. See this information in framework: -Compression-tutoriaI-tp30554.html Sent from the Point Cloud Library (PCL) maiIing list mailing Iist save at Nabble.com. / /.