vlogroup / dvs-reconstruction Goto Github PK

I've successfully built the dvs-reconstruction, and also I'm able to run offline_reconstruction on the CPU. At times when the .aer2 file is of a huge chunk it either runs out of memory or hangs the whole system. Due to which I was trying it to run it on GPU with the flag -g. My command looks like ./offline_reconstruction -f "event-based.aer2 file" -o "output_folder" -g
The command throws an error:

terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<boost::program_options::invalid_command_line_syntax> >' what(): the required argument for option '--gpu' is missing Aborted (core dumped)

In the final make, I'm getting a lot (probably a couple dozen) of undefined reference to boost::program_options::...` flags and the recipe for target 'offline_reconstruction' failed.

I thought perhaps it was an issue with boost, but when looking in the cmake ../src output, I do see

-- Boost version: 1.58.0
-- Found the following Boost libraries:
--   program_options
--   filesystem
--   system

Is there a minimum version for Boost? Or do I need to check elsewhere for the source of the errors?

Which version of libcaer driver did you use? The link you provided is now incorrec, can you put the driver version you used in the project?
Thank you very much!

I followed the step and I can build it. but i got below error:

git clone --recursive https://github.com/VLOGroup/dvs-reconstruction

cd cnpy
cmake .
make
sudo make install
cd ../libcaer
cmake .
make
sudo make install

i got:
..
..
-- Up-to-date: /usr/local/lib/libcaer.so.2
-- Up-to-date: /usr/local/lib/libcaer.so

cd ..
mkdir build
cd build
cmake ../src
make -j6

But when i tried to run the bin, i got below error.

raymond@ubuntu1:~/dvs/dvs-reconstruction/build$ ./live_reconstruction_gui
./live_reconstruction_gui: error while loading shared libraries: libcaer.so.2: cannot open shared object file: No such file or directory

I guess i should not use sudo install?? No idea on how to fix the problem now..
Thanks for advice.

Hi, I am starting to look into neuromorphic computing and needed to know that
-> Whether the code work with the images form normal camera, or only need event data to reconstruct the intensity images.

-> I have the dat files containing event data . How and where should I put the files to run the code ?

-> Do I need libcaer library, if I am not using the camera.

P.S. Sorry for posting this here, and if the question sounds silly, I am just starting.

With regards,
Sameer Purwar

Hi,

I followed the step to install imageutilities.
$ cd build
$ cmake ..
$ make
$ make install
(https://github.com/VLOGroup/imageutilities)

I set the IMAGEUTILITIES_ROOT to
/home/lscm/dvs/dvs-reconstruction/imageutilities

I installed cnpy,libcaer.

Then I ran below line and I got error message:
~/dvs/dvs-reconstruction/build$ cmake ../src
-- Imageutilities requested components: iucore;iuio;iugui;iumath
CMake Error at /home/lscm/dvs/dvs-reconstruction/imageutilities/ImageUtilitiesConfig.cmake:137 (message):
iuio marked as required but not found
Call Stack (most recent call first):
CMakeLists.txt:25 (find_package)

-- Configuring incomplete, errors occurred!

Thanks.

Hi, I would like to test your algorithm. So, I've succeed with building, with some minor problems. But after I have tried to run it on some sample camera log, I've found that it requires card compute capability 3.0.

I have a pretty old laptop (check below).

As you for sure know your codebase better than me (I found your article/repo today), e.g. what exact 3.0 features it uses, etc.

So, I would like to ask, if it makes sense to even try to patch/replace code depending on -arch=sm_30?

• https://github.com/VLOGroup/imageutilities
• https://github.com/VLOGroup/dvs-reconstruction

Or it would be too much work, and I should go and find some appropriate hardware for tests?

Thank you.

./deviceQuery Starting...

 CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 1 CUDA Capable device(s)

Device 0: "Quadro 1000M"
  CUDA Driver Version / Runtime Version          7.5 / 7.5
  CUDA Capability Major/Minor version number:    2.1
  Total amount of global memory:                 2047 MBytes (2146631680 bytes)
  ( 2) Multiprocessors, ( 48) CUDA Cores/MP:     96 CUDA Cores
  GPU Max Clock rate:                            1400 MHz (1.40 GHz)
  Memory Clock rate:                             900 Mhz
  Memory Bus Width:                              128-bit
  L2 Cache Size:                                 131072 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(65536), 2D=(65536, 65535), 3D=(2048, 2048, 2048)
  Maximum Layered 1D Texture Size, (num) layers  1D=(16384), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(16384, 16384), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 32768
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  1536
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (65535, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 1 copy engine(s)
  Run time limit on kernels:                     Yes
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device PCI Domain ID / Bus ID / location ID:   0 / 1 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 7.5, CUDA Runtime Version = 7.5, NumDevs = 1, Device0 = Quadro 1000M
Result = PASS

I can open the saved data like this:

20.0405 32 115 -1
20.0405 32 115 -1
20.0405 32 114 -1
20.0405 33 115 -1
20.0406 25 113 -1
20.0406 23 121 1
20.0406 15 115 1
20.0406 16 113 1
20.0406 15 114 1
20.0406 17 119 1

is that a good way to isolate the data sequence when an object is seen by the cam?

I wonder if the event/image = 1000 means a moving average of 1000 to build the reconstruction image?

thanks.

Hi,
I build and install all dependencies and finally build src as well, according to the README. I encountered the same problem as #8. I am unable to launch the live_reconstruction_gui. It shows the following error:

QXcbIntegration: Cannot create platform OpenGL context, neither GLX nor EGL are enabled
QXcbIntegration: Cannot create platform OpenGL context, neither GLX nor EGL are enabled
QOpenGLWidget: Failed to create context
QXcbIntegration: Cannot create platform OpenGL context, neither GLX nor EGL are enabled
QOpenGLWidget: Failed to create context
QXcbIntegration: Cannot create platform OpenGL context, neither GLX nor EGL are enabled
QOpenGLWidget: Failed to create context
composeAndFlush: makeCurrent() failed

We use NVIDIA-SMI 410.78 on Ubuntu 16.04. I'm using Qt version 5.12.0 and CUDA version 10.0.
We are working on the same problem and wanted to compare results. I learned from #8 that you have published an offline evaluation script. I also generated the file of'offline_reconstruction_XXX' in the build folder after following the README operation.

I would like to ask you how to input a dataset and run the offline evaluation script to generate reconstructed event images?
Could you help me resolve this issue? Thanks

Hi there!

thank you very much for sharing this code. I am Federico and I work at IniLabs GmbH, we produce and sell the neuromorphic vision sensors.

I am now trying to compile your code, however I am getting the following error, do you have any ideas on how to solve this?

thank you very much for your support. I look forward seeing the intensity reconstruction in live action.

attached the log

[ 6%] Automatic moc for target live_reconstruction_gui
[ 6%] Built target live_reconstruction_gui_automoc
[ 13%] Automatic moc for target common
[ 13%] Built target common_automoc
[ 46%] Built target common
[ 53%] Linking CXX executable live_reconstruction_gui
Running ld for x86_64 ...
Undefined symbols for architecture x86_64:
"iu::Qt5ImageGpuWidget::Qt5ImageGpuWidget(iu::Size<2u>, QWidget*)", referenced from:
ReconstructionMainWindow::ReconstructionMainWindow(QWidget*, std::__1::vector<Event, std::__1::allocator >&) in reconstructionmainwindow.cpp.o
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make[2]: *** [live_reconstruction_gui] Error 1
make[1]: *** [CMakeFiles/live_reconstruction_gui.dir/all] Error 2
make: *** [all] Error 2

Hi,
I build and install all dependencies and finally build src as well, according to the README. But I am unable to launch the live_reconstruction_gui. It shows the following error:

QXcbIntegration: Cannot create platform OpenGL context, neither GLX nor EGL are enabled
QXcbIntegration: Cannot create platform OpenGL context, neither GLX nor EGL are enabled
QOpenGLWidget: Failed to create context
QXcbIntegration: Cannot create platform OpenGL context, neither GLX nor EGL are enabled
QOpenGLWidget: Failed to create context
QXcbIntegration: Cannot create platform OpenGL context, neither GLX nor EGL are enabled
QOpenGLWidget: Failed to create context
composeAndFlush: makeCurrent() failed

What am I missing? Could you help me resolve this issue? Thanks

Hi,

I got a Cuda error when running the GUI. After I load the sample data and hit play, the GUI terminated. Here is everything I got from the terminal after I run live_reconstruction_gui:

XmbTextListToTextProperty result code -2
XmbTextListToTextProperty result code -2
CUDA error: cudaErrorUnknown:unknown error
File: /home/SENSETIME/wangdaoye/PycharmProjects/event-deblur/others_code/dvs-reconstruction/imageutilities/src/iugui/qt5imagegpuwidget.cpp
Function: init_cuda
Line: 163
CUDA error: cudaErrorUnknown:unknown error
File: /home/SENSETIME/wangdaoye/PycharmProjects/event-deblur/others_code/dvs-reconstruction/imageutilities/src/iugui/qt5imagegpuwidget.cpp
Function: init_cuda
Line: 163
CUDA error: cudaErrorUnknown:unknown error
File: /home/SENSETIME/wangdaoye/PycharmProjects/event-deblur/others_code/dvs-reconstruction/imageutilities/src/iugui/qt5imagegpuwidget.cpp
Function: init_cuda
Line: 163
OpenGL profile: Core
OpenGL version: 3.3
Swap behaviour: 2
Render type: 1
XmbTextListToTextProperty result code -2
XmbTextListToTextProperty result code -2
XmbTextListToTextProperty result code -2
XmbTextListToTextProperty result code -2
XmbTextListToTextProperty result code -2
XmbTextListToTextProperty result code -2
terminate called after throwing an instance of 'IuCudaException'
what(): IuException: CUDA Error: unknown error
where: /home/SENSETIME/wangdaoye/PycharmProjects/event-deblur/others_code/dvs-reconstruction/imageutilities/src/iucore/image_allocator_gpu.h | copy:34
Aborted (core dumped)

I wonder what's going wrong. Thanks!!

Hi, I was able to compile, but when I try running live_reconstruction_gui, it throws Segmentation fault (core dumped) error. I appreciate if you could help. Thanks.

I have successfully installed the offline reconstruction as well as the live reconstruction gui ... I ran the offline reconstruction code with sample data (aer2 files) that were provided. It gave me a successful reconstruction... How do I use the same for event data which I have in the form of .bag files?.. Directly passing it as -f argument doesn't work... Thanks

Hi, after building with the WITH_DAVIS set to ON, I tested offline on a small sequence but the output is 128x128 instead of 240x180.

HI,

I spended few weeks but still I cannot install IMAGEUTILITIES.
I decide to start everything from scratch again.
Can you recommend the ubuntu version, openCV version, cuda version, that proved be working config for install this project?

Thanks.