Add RSSDK 2.0 (librealsense2) grabber #2214

daniel-packard · 2018-02-13T20:52:03Z

Hi PCL team - @patrickabadi has implemented a librealsense 2.0 based RealSense2Grabber, and we'd like to make a contribution. I noticed the last comment on #1633 implies @UnaNancyOwen was planning to implement a librealsense 2.0 grabber, but maybe this will provide a good starting point that covers a lot of the boilerplate cmake configuration.

quick notes:

we have only tested on windows 10
we have only tested with v2.10 of librealsense
if you try to use librealsense with Qt5, you will be affected by the macro issue reported here: QT Linking Issues IntelRealSense/librealsense#1059

cmake/Modules/FindLibRealSense.cmake

daniel-packard · 2018-02-13T21:04:45Z

io/CMakeLists.txt

@@ -5,9 +5,9 @@ set(SUBSYS_DEPS common octree)
 set(build TRUE)
 PCL_SUBSYS_OPTION(build "${SUBSYS_NAME}" "${SUBSYS_DESC}" ON)
 if(WIN32)
-    PCL_SUBSYS_DEPEND(build "${SUBSYS_NAME}" DEPS ${SUBSYS_DEPS} OPT_DEPS openni openni2 ensenso davidSDK dssdk rssdk pcap png vtk)
+    PCL_SUBSYS_DEPEND(build "${SUBSYS_NAME}" DEPS ${SUBSYS_DEPS} OPT_DEPS openni openni2 ensenso davidSDK dssdk rssdk librealsense pcap png vtk)


I'm assuming that since librealsense is cross platform it should be added to both of these lists

SergioRAgostinho · 2018-02-13T22:17:24Z

Patrick and Daniel thanks for contributing this! :)

We're still in C++03 and the migration to C++14 is bound to only happen after our next release, meaning this is somewhat low priority for now. I hope that doesn't conflict too much with your "agenda".

daniel-packard · 2018-02-14T00:02:46Z

Thanks sergio - since we are working from PCL+librealsense source it's working fine for us. There is no rush to merge on our side.

Just let us know when the C++ 11 migration is live, and we'll bring this branch up to date with PCL master

UnaNancyOwen · 2018-02-14T14:15:11Z

io/include/pcl/io/real_sense_2_grabber.h

+	class PCL_EXPORTS RealSense2Grabber : public pcl::Grabber
+	{
+	public:
+		RealSense2Grabber();


I think it would be good so that to receive the serial number in constructor to open specific devices or multiple devices.

public: - RealSense2Grabber (); + RealSense2Grabber (const std::string& serial_number = ""); protected: + std::string serial_number;

- RealSense2Grabber::RealSense2Grabber () + RealSense2Grabber::RealSense2Grabber (const std::string& serial_number) : running (false) , quit (false) + , serial_number (serial_number)

void RealSense2Grabber::start () { running = true; rs2::config cfg; + if (!serial_number.empty()) + cfg.enable_device (serial_number);

You can now start the grabber with a specific device

UnaNancyOwen · 2018-02-14T14:17:56Z

io/src/real_sense_2_grabber.cpp

+    auto uvptr = points.get_texture_coordinates();
+
+    uint8_t r, g, b;
+    for (auto& p : cloud->points)


Please consider to parallelizing using OpenMP.

I've added OpenMP to the loops. Please let me know if I'm using the standard PCL OpenMP conventions.

UnaNancyOwen · 2018-02-14T14:24:29Z

io/src/real_sense_2_grabber.cpp

+
+	cfg.enable_stream(RS2_STREAM_COLOR, 424, 240, RS2_FORMAT_RGB8, 30);
+	cfg.enable_stream(RS2_STREAM_DEPTH, 424, 240, RS2_FORMAT_ANY, 30);
+	cfg.enable_stream(RS2_STREAM_INFRARED, 424, 240, RS2_FORMAT_ANY, 30);


I think that it should provide some way to set resolution and frame rate.

This can now be updated:

https://github.com/complexrehabsystems/pcl/blob/e342155a33a8a9aa2cf3322e8c74782e8d88b6c3/io/src/real_sense_2_grabber.cpp#L51

io/src/real_sense_2_grabber.cpp

UnaNancyOwen · 2018-02-14T14:43:07Z

cmake/Modules/FindLibRealSense.cmake

+
+# Libraries
+find_library(LIBREALSENSE_LIBRARY NAMES librealsense.lib)
+find_library(LIBREALSENSE_LIBRARY_DEBUG NAMES librealsense.lib)


I think the librealsense2 library name is realsense2.lib.
Why is it librealsense2.lib?

And, Please add suffix (e.g. _d) for debug library name. (realsense2_d.lib)

UnaNancyOwen · 2018-02-14T14:50:05Z

@SergioRAgostinho Where is used C++11 features?

UnaNancyOwen · 2018-02-14T15:23:48Z

cmake/Modules/FindLibRealSense.cmake

+find_package(PkgConfig QUIET)
+
+# Include directories
+find_path(LIBREALSENSE_INCLUDE_DIR librealsense2/rs.h)


Please consider to add these options (HINTS, PATHS, and PATH__SUFFIXES) to find_path() and find_library().

I think it can write as following.
(But, I'm not familiar with Linux, and MacOS. It need more comments from other reviewer.)

find_path(LIBREALSENSE_INCLUDE_DIR librealsense2/rs.h PATHS "$ENV{realsense2_DIR}" "$ENV{PROGRAMW6432}/librealsense2" # for self built library "$ENV{PROGRAMFILES}/librealsense2" # for self built library "$ENV{PROGRAMFILES}/Intel RealSense SDK 2.0" # for pre built library # "Please specify search paths for Linux and MacOS" PATH_SUFFIXES include) set(LIBREALSENSE_INCLUDE_DIRS ${LIBREALSENSE_INCLUDE_DIR}) mark_as_advanced(LIBREALSENSE_INCLUDE_DIRS) set(LIBREALSENSE_RELEASE_NAME realsense2) set(LIBREALSENSE_DEBUG_NAME realsense2_d) set(LIBREALSENSE_SUFFIX x86) if(CMAKE_SIZEOF_VOID_P EQUAL 8) set(LIBREALSENSE_SUFFIX x64) endif() find_library(LIBREALSENSE_LIBRARY NAMES ${LIBREALSENSE_RELEASE_NAME} PATHS "$ENV{realsense2_DIR}" "$ENV{PROGRAMW6432}/librealsense2" "$ENV{PROGRAMFILES}/librealsense2" "$ENV{PROGRAMFILES}/Intel RealSense SDK 2.0" # "Please specify search paths for Linux and MacOS" PATH_SUFFIXES lib lib/${LIBREALSENSE_SUFFIX}) find_library(LIBREALSENSE_LIBRARY_DEBUG NAMES ${LIBREALSENSE_DEBUG_NAME} PATHS "$ENV{realsense2_DIR}" "$ENV{PROGRAMW6432}/librealsense2" "$ENV{PROGRAMFILES}/librealsense2" "$ENV{PROGRAMFILES}/Intel RealSense SDK 2.0" # "Please specify search paths for Linux and MacOS" PATH_SUFFIXES lib lib/${LIBREALSENSE_SUFFIX}) if(NOT LIBREALSENSE_LIBRARY_DEBUG) set(LIBREALSENSE_LIBRARY_DEBUG ${LIBREALSENSE_LIBRARY}) endif() set(LIBREALSENSE_LIBRARIES optimized ${LIBREALSENSE_LIBRARY} debug ${LIBREALSENSE_LIBRARY_DEBUG}) mark_as_advanced(LIBREALSENSE_LIBRARY LIBREALSENSE_LIBRARY_DEBUG)

UnaNancyOwen · 2018-02-14T15:37:46Z

@daniel-packard Thank you for your contribution.
I did the first review. Please tell us your opinion.
(I had not started implementing grabber yet. Thanks 👍)

SergioRAgostinho · 2018-02-14T15:46:13Z

Where is used C++11 features?

https://github.com/IntelRealSense/librealsense/releases/tag/v2.8.0
First official build mentions C++11 support. I assume the library uses C++11 features.

In the actual code
Use of a for-range loop : https://github.com/PointCloudLibrary/pcl/pull/2214/files#diff-2c7a8b41ca6e6bcff66df4b94d4d76b0R108
Use of auto type: https://github.com/PointCloudLibrary/pcl/pull/2214/files#diff-2c7a8b41ca6e6bcff66df4b94d4d76b0R132
Use of tupples: https://github.com/PointCloudLibrary/pcl/pull/2214/files#diff-2c7a8b41ca6e6bcff66df4b94d4d76b0R148

Guys, some really really quick (review) comments:

since you're already dealing with C++11, you can start using std::thread and std::mutex.
Don't bother calling lock.unlock () on unique locks. Supposedly they do that automatically when they run out of scope.
There's no docstrings anywhere, but I can see you might want to do these later.
There's no unit tests, although for this we're gonna need to mock a grabber. I might help you guys set up some boiler plate code later on since AFAIK there's nothing of the sort in the library so far.

UnaNancyOwen · 2018-02-14T16:06:52Z

io/src/real_sense_2_grabber.cpp

+    return 30.0f;
+}
+
+pcl::PointCloud<pcl::PointXYZ>::Ptr RealSense2Grabber::convertDepthToPointXYZ(const rs2::points & points)


Not implemented yet?

This is now implemented

daniel-packard · 2018-02-14T16:52:21Z

cmake/Modules/FindLibRealSense.cmake

+
+# Libraries
+set(LIBREALSENSE_RELEASE_NAME realsense2)
+set(LIBREALSENSE_DEBUG_NAME realsense2_d)


Of course, you are right about the name of the libraries (librealsense.lib vs realsense2) 👍

I also liked your recommended update to include PATHS and PATH_SUFFIXES in the calls to find_path and find_library, so I've updated that as well. I don't know what appropriate values would be on linux/mac either -- hopefully a Linux/Mac user can comment with additional recommendations for those platforms

patrickabadi · 2018-02-14T23:24:27Z

@SergioRAgostinho I've implemented some of the C++11 changes you mentioned regarding std::thread and std::mutex.

You mentioned help in setting up unit tests for the grabber. Yeah that would definitely be helpful but we don't want the scope of this pull request to get too big.

UnaNancyOwen · 2018-02-15T04:59:48Z

@daniel-packard I'd like to adjust coding style according to PCL.
Your source code will become better. It makes it easier for everyone to read and review.
Thanks for your cooperation.

Please adjust indent of whole source code. (change to two space from tab.)

- 	int32_t i;
+   int32_t i;

Please add one space before call function.

- function();
+ function ();

Please new line after function return type.

- int32_t function ();
+ int32_t
+ function ();

~~Please change to cstdint types. (e.g. int32_t)~~

- int i;
+ int32_t i;

SergioRAgostinho · 2018-02-15T09:38:44Z

You mentioned help in setting up unit tests for the grabber. Yeah that would definitely be helpful but we don't want the scope of this pull request to get too big.

What I mentioned was setting up boilerplate code to exemplify how to mock a grabber. The unit tests and the actual mocking behavior will need to be written by you guys because you have the hardware so actually know how to properly mock it. Providing unit tests is well within the scope of this PR.

Please change to cstdint types.

I don't usually request this unless there's some compelling argument behind it. I do normally advise to be considerate about your variable types based on the prior information i.e.:

if you know the max range in advance then sure, spec it for a certain number of bits.
use size_t by default for specifying indices and container sizes, unless you're sure the range will be smaller somehow.
Use the right types for the right job e.g. don't use a signed integer to count things since a count can't be a negative number.

@UnaNancyOwen any specific motives behind this request I'm failing to see?

SergioRAgostinho · 2018-02-15T10:27:03Z

.gitignore

@@ -0,0 +1 @@
+build/


Remove this please. Add it to your .git/info/exclude file.

You just taught me a new git trick - thanks!

And it is removed.

UnaNancyOwen · 2018-02-15T10:37:13Z

I do normally advise to be considerate about your variable types based on the prior information i.e.:

if you know the max range in advance then sure, spec it for a certain number of bits.

use size_t by default for specifying indices and container sizes, unless you're sure the range will be smaller somehow.

Use the right types for the right job e.g. don't use a signed integer to count things since a count can't be a negative number.

@SergioRAgostinho I was intended for this. Sorry, ｍy request expression was not appropriate.

UnaNancyOwen · 2018-02-15T14:05:50Z

It is an first idea for support capture from bag file.
What do you think?

class PCL_EXPORTS RealSense2Grabber : public pcl::Grabber
{
  public:
    RealSense2Grabber (const uint32_t serial_number = 0);
    RealSense2Grabber (const std::string& file_name);
  protected:
    uint32_t serial_number;
    std::string file_name;
}

// capture from sensor
RealSense2Grabber::RealSense2Grabber (const uint32_t serial_number = 0)
: serial_number (serial_number)
{
}

// capture from file
RealSense2Grabber (const std::string& file_name)
: file_name (file_name)
{
}

void
RealSense2Grabber::start ()
{
  running = true;

  rs2::config cfg;

  // capture from file
  if (file_name.rfind (".bag") == 4)
  {
    cfg.enable_device_from_file (file_name);
  }
  // capture from sensor
  else
  {
    if (serial_number)
      cfg.enable_device (std::to_string (serial_number));

    cfg.enable_stream (RS2_STREAM_COLOR, deviceWidth, deviceHeight, RS2_FORMAT_RGB8, targetFps);
    cfg.enable_stream (RS2_STREAM_DEPTH, deviceWidth, deviceHeight, RS2_FORMAT_ANY, targetFps);
    cfg.enable_stream (RS2_STREAM_INFRARED, deviceWidth, deviceHeight, RS2_FORMAT_ANY, targetFps);
  }

  pipe.start(cfg);

  thread = std::thread(&RealSense2Grabber::threadFunction, this);
}

daniel-packard · 2018-02-15T15:11:25Z

@daniel-packard I'd like to adjust coding style according to PCL.

A quick comment that it is actually @patrickabadi who has contributed the C++ code, so all praise/criticism should be directed towards him 😁

That said, we are happy to reformat according to the PCL style guide

http://www.pointclouds.org/documentation/advanced/pcl_style_guide.php

Have you considered adding a .editorconfig file to the repo to help facilitate some of the auto-formatter settings? (e.g. https://gist.github.com/buoto/7e2cff6c17c108658fc751df845b1040)

What I mentioned was setting up boilerplate code to exemplify how to mock a grabber. [...] Providing unit tests is well within the scope of this PR.

Ah, I misunderstood what was being offered/requested -- yes please, your help setting this up would be very appreciated, and we can provide unit tests (although we might require some guidance, or pointers other implementations within PCL that would provide a good model )

It is an first idea for support capture from bag file. What do you think?

Isn't a bag file a concept from ROS, or has PCL adopted it too? In general, I think the ability to support record/playback is excellent -- and if this is the recommended way, we will work together to add support

Last note -- we are amazed at the engagement from you guys, and the excellent feedback 💯

We will have to start spreading out our contributions a little bit so we can continue working on our day jobs too. What is the approximate timeframe for adding C++ 11 support so we can make our plans accordingly?

UnaNancyOwen · 2018-02-15T15:22:11Z

Isn't a bag file a concept from ROS, or has PCL adopted it too? In general, I think the ability to support record/playback is excellent -- and if this is the recommended way, we will work together to add support

librealsense2 have feature that capture from bag files. It is easy and simple.
I think it is good to support playback in this grabber.
https://github.com/IntelRealSense/librealsense/blob/ba01147d65db16fdf4da36a3e718fe81c8421034/doc/sample-data.md

rs2::config config;
config.enable_device_from_file ("file.bag");
rs2::pipeline pipeline;
pipeline.start (config);

patrickabadi · 2018-02-15T19:54:39Z

@UnaNancyOwen I updated the code to match the pcl style guide. Please let me know if everything looks good. I also added the bag file loading capability.

Note: We've been testing RealSense2Grabber on the pcl in_hand_scanner sample and so far it was been working great.

UnaNancyOwen · 2018-02-16T13:54:08Z

io/include/pcl/io/real_sense_2_grabber.h

+    /** \brief Constructor
+    * \param[in] serial_number optional to start up a specific device
+    */
+    RealSense2Grabber (const std::string& serial_number = "", const std::string& file_name = "");


In this implementation, It need to pass two arguments when capturing from file.
I think that this specification is generally not favored.

// capture from bag file boost::shared_ptr<pcl::RealSense2Grabber> grabber = boost::make_shared<pcl::RealSense2Grabber> ("", "file.bag");

Is there a special reason for this implementation?
I think it should define two contractors. (#2214 (comment))

Hi @UnaNancyOwen that makes sense since the user can only do one or the other. I've separated the constructors now.

UnaNancyOwen · 2018-02-16T14:51:13Z

cmake/Modules/FindLibRealSense.cmake

+                "$ENV{PROGRAMFILES}/librealsense2" # for self built library
+                "$ENV{PROGRAMFILES}/Intel RealSense SDK 2.0" # for pre built library
+                "$ENV{LIBREALSENSE_ROOT}"
+                # "Please specify search paths for Linux and MacOS"


Should it add "/usr/local/" and "/usr/" for Linux and MacOS?

Seems reasonable, and looking at libusb, these seem to be some of the standard hints on linux: https://github.com/PointCloudLibrary/pcl/blob/46cb8fe5589e88e36d79f9b8b8e5f4ff4fceb5de/cmake/Modules/Findlibusb-1.0.cmake

I'll do a quick review of the librealsense docs and see if they install to any of these paths by default on linux/mac

UnaNancyOwen · 2019-06-12T08:27:24Z

I have RealSense D415, D435, and D435i. I will test with them, and test with ros-bag file on Windows.
And, I will wait for the taketwo's report. 👍

cmake/Modules/FindRSSDK2.cmake

UnaNancyOwen · 2019-06-12T14:45:23Z

PCLConfig.cmake.in

@@ -216,6 +216,19 @@ macro(find_rssdk)
  find_package(RSSDK)
 endmacro()

+macro(find_rssdk2)
+


Please remove extra blank line in first and last of this macro.

UnaNancyOwen · 2019-06-12T15:51:45Z

io/include/pcl/io/real_sense_2_grabber.h

+ */
+
+#ifndef PCL_IO_REAL_SENSE_2_GRABBER_H
+#define PCL_IO_REAL_SENSE_2_GRABBER_H


We replaced all include-guards to #pragma once in #2617.
Please replace this include-guard.

- #ifndef PCL_IO_REAL_SENSE_2_GRABBER_H - #define PCL_IO_REAL_SENSE_2_GRABBER_H + #pragma once ... - #endif /* PCL_IO_REAL_SENSE_2_GRABBER_H */

UnaNancyOwen · 2019-06-12T15:56:10Z

I have confirmed that this grabber works correctly with sensor and file. 👍
Please consider my three comments. Thanks,

taketwo

Just a few inline comments, plus one bigger question. Could you please explain why the critical sections are where they are? E.g. why mutex needs to be locked while busy-waiting for frames?

taketwo · 2019-06-13T19:51:37Z

PCLConfig.cmake.in

+  elseif(NOT realsense2_DIR)
+    get_filename_component(realsense2_DIR "@REALSENSE2_INCLUDE_DIRS@" PATH)
+    set(realsense2_DIR "${realsense2_DIR}/lib/cmake/realsense2" CACHE PATH "The directory containing realsense2Config.cmake")
+  endif(PCL_ALL_IN_ONE_INSTALLER)


This is an outdated CMake syntax to repeat the if condition in endif and in the meantime we got rid of this everywhere in PCL. Please remove here as well.

taketwo · 2019-06-13T19:52:15Z

PCLConfig.cmake.in

+    set(realsense2_DIR "${realsense2_DIR}/lib/cmake/realsense2" CACHE PATH "The directory containing realsense2Config.cmake")
+  endif(PCL_ALL_IN_ONE_INSTALLER)
+  find_package(RSSDK2)
+endmacro(find_rssdk2)


taketwo · 2019-06-13T20:05:35Z

io/src/real_sense_2_grabber.cpp

+      cfg.enable_stream (RS2_STREAM_DEPTH, device_width_, device_height_, RS2_FORMAT_ANY, target_fps_);
+      cfg.enable_stream (RS2_STREAM_INFRARED, device_width_, device_height_, RS2_FORMAT_ANY, target_fps_);
+    }    
+


Yep, I'm still in favor of the last Tsukasa's proposal.

patrickabadi · 2019-06-13T21:04:37Z

Just a few inline comments, plus one bigger question. Could you please explain why the critical sections are where they are? E.g. why mutex needs to be locked while busy-waiting for frames?

I just looked at it and I think it might be old code, there's no reason to lock while getting data. I'll take it out

taketwo · 2019-06-14T10:03:40Z

io/src/real_sense_2_grabber.cpp

+    {
+      sw.reset ();
+
+      rs2::depth_frame depth = nullptr;


This fails to compile with GCC 7.4 on Ubuntu.

error: conversion from ‘std::nullptr_t’ to non-scalar type ‘rs2::depth_frame’ requested

Which makes sense, how can you assign nullptr to a class?

taketwo · 2019-06-16T15:33:20Z

I've added librealsense to our Ubuntu 16.04 image that is used on CI. From now on we can see whether the new grabber can be built on that platform. I've also restarted the previous run, so you can see the error I poster before.

taketwo · 2019-06-19T07:55:32Z

I can successfully build the grabber now. However, when I try to visualize point clouds from the grabber, I get:

terminate called after throwing an instance of 'std::runtime_error'
  what():  Profile does not contain the requested stream
Aborted (core dumped)

I'm using D415 and setDeviceOptions(640, 480, 30). Did anyone had any luck with using this on Linux?

daniel-packard · 2019-06-23T21:53:09Z

Did anyone had any luck with using this on Linux?

I haven't had a chance to try in linux.

That error comes from librealsense. Are you able to use the D415 with the "RealSense Viewer" that is provided by librealsense?

Will try to spin up an Ubuntu VM to test

taketwo · 2019-06-24T07:21:06Z

Yep, I tried with their viewer and it worked. Also different combinations of resolution/fps.

SergioRAgostinho · 2019-11-11T10:48:43Z

Sorry to bring this from the dead.

I would be interested in releasing the working Windows version of this for this release, even if under some <pcl/experimental/.h> header or under a pcl::experimental namespace. At least the functionality we know it is working. @PointCloudLibrary/maintainers does this sound reasonable?

taketwo · 2019-11-11T10:58:54Z

Fine for me. I don't think we need to use "experimental" directory, it's sufficient to mention this explicitly in the documentation.

SergioRAgostinho · 2019-11-11T12:10:02Z

OK. My plan was to compile and include these files only under Windows and fork a new a issue to address the failures under Linux.

SergioRAgostinho · 2019-11-12T08:34:14Z

io/src/real_sense_2_grabber.cpp

+
+    if (signal_PointXYZRGB->num_slots () > 0 || signal_PointXYZRGBA->num_slots () > 0)
+    {
+      assert ( prof.get_stream ( RS2_STREAM_COLOR ).format () == RS2_FORMAT_RGB8 );


This check is irrelevant because this condition is already tested in line 129. In case it is false an exception is thrown.

SergioRAgostinho · 2019-11-12T08:37:37Z

io/src/real_sense_2_grabber.cpp

+      assert ( prof.get_stream ( RS2_STREAM_COLOR ).format () == RS2_FORMAT_RGB8 );
+    }
+
+    assert ( prof.get_stream ( RS2_STREAM_DEPTH ).format () == RS2_FORMAT_Z16 );


Same comment as above, the check is already performed in line 130.

SergioRAgostinho · 2019-11-13T07:14:44Z

@taketwo, @UnaNancyOwen, @kunaltyagi Do you mind if I merge this and make the corrections in another PR?

UnaNancyOwen · 2019-11-13T07:27:36Z

I agree.
This pull request has been open for too long. We should not continue to fix here.
I think it is good to open a new pull request after merged this pull request.

daniel-packard · 2019-11-14T16:32:54Z

Thanks everyone! 🎉

daniel-packard commented Feb 13, 2018

View reviewed changes

SergioRAgostinho added c++11 needs: code review Specify why not closed/merged yet labels Feb 13, 2018

SergioRAgostinho added the module: io label Feb 14, 2018

UnaNancyOwen reviewed Feb 14, 2018

View reviewed changes

io/src/real_sense_2_grabber.cpp Show resolved Hide resolved

UnaNancyOwen reviewed Feb 14, 2018

View reviewed changes

SergioRAgostinho added needs: author reply Specify why not closed/merged yet and removed needs: code review Specify why not closed/merged yet labels Feb 14, 2018

UnaNancyOwen reviewed Feb 14, 2018

View reviewed changes

daniel-packard commented Feb 14, 2018

View reviewed changes

SergioRAgostinho reviewed Feb 15, 2018

View reviewed changes

UnaNancyOwen reviewed Feb 16, 2018

View reviewed changes

UnaNancyOwen reviewed Jun 12, 2019

View reviewed changes

cmake/Modules/FindRSSDK2.cmake Show resolved Hide resolved

UnaNancyOwen reviewed Jun 12, 2019

View reviewed changes

Add RSSDK2_FOUND message, replace ifndef include guard with pragma once

69ba109

taketwo reviewed Jun 13, 2019

View reviewed changes

patrickabadi and others added 3 commits June 13, 2019 14:13

Remove mutex lock. Not needed

65fc64d

exception to ensure proper stream format

131775c

Remove redundant closing if(XYZ) conditional text

b5a5e9c

taketwo reviewed Jun 14, 2019

View reviewed changes

Remove forward declarations of rgb, ir, only create on needed

36d3653

SergioRAgostinho reviewed Nov 12, 2019

View reviewed changes

SergioRAgostinho merged commit bc898b6 into PointCloudLibrary:master Nov 14, 2019

SergioRAgostinho mentioned this pull request Nov 14, 2019

Changes to RSSDK 2.0 grabber #3467

Closed

3 tasks

taketwo added changelog: new feature Meta-information for changelog generation and removed c++14 labels Jan 14, 2020

taketwo changed the title ~~Add RSSDK 2.0 (librealsense 2.0) grabber implementation~~ Add RSSDK 2.0 (librealsense2) grabber Jan 18, 2020

Zhangxh251 mentioned this pull request Apr 3, 2023

[compile error] "Compile PCL with librealsense support" #5427

Open

		@@ -0,0 +1 @@
		build/

Add RSSDK 2.0 (librealsense2) grabber #2214

Add RSSDK 2.0 (librealsense2) grabber #2214

Conversation

daniel-packard commented Feb 13, 2018 • edited Loading

Choose a reason for hiding this comment

SergioRAgostinho commented Feb 13, 2018

daniel-packard commented Feb 14, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

UnaNancyOwen Feb 14, 2018 • edited Loading

Choose a reason for hiding this comment

UnaNancyOwen commented Feb 14, 2018

UnaNancyOwen Feb 14, 2018 • edited Loading

Choose a reason for hiding this comment

UnaNancyOwen commented Feb 14, 2018

SergioRAgostinho commented Feb 14, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

daniel-packard Feb 14, 2018 • edited Loading

Choose a reason for hiding this comment

patrickabadi commented Feb 14, 2018

UnaNancyOwen commented Feb 15, 2018 • edited Loading

SergioRAgostinho commented Feb 15, 2018

SergioRAgostinho Feb 15, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

UnaNancyOwen commented Feb 15, 2018

UnaNancyOwen commented Feb 15, 2018 • edited Loading

daniel-packard commented Feb 15, 2018

UnaNancyOwen commented Feb 15, 2018 • edited Loading

patrickabadi commented Feb 15, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

UnaNancyOwen commented Jun 12, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

UnaNancyOwen commented Jun 12, 2019

taketwo left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

patrickabadi commented Jun 13, 2019

taketwo Jun 14, 2019 • edited Loading

Choose a reason for hiding this comment

taketwo commented Jun 16, 2019

taketwo commented Jun 19, 2019

daniel-packard commented Jun 23, 2019

taketwo commented Jun 24, 2019

SergioRAgostinho commented Nov 11, 2019 • edited Loading

taketwo commented Nov 11, 2019

SergioRAgostinho commented Nov 11, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

SergioRAgostinho commented Nov 13, 2019

UnaNancyOwen commented Nov 13, 2019

daniel-packard commented Nov 14, 2019

daniel-packard commented Feb 13, 2018 •

edited

Loading

UnaNancyOwen Feb 14, 2018 •

edited

Loading

UnaNancyOwen Feb 14, 2018 •

edited

Loading

daniel-packard Feb 14, 2018 •

edited

Loading

UnaNancyOwen commented Feb 15, 2018 •

edited

Loading

SergioRAgostinho Feb 15, 2018 •

edited

Loading

UnaNancyOwen commented Feb 15, 2018 •

edited

Loading

UnaNancyOwen commented Feb 15, 2018 •

edited

Loading

UnaNancyOwen commented Jun 12, 2019 •

edited

Loading

taketwo Jun 14, 2019 •

edited

Loading

SergioRAgostinho commented Nov 11, 2019 •

edited

Loading