Hi,

I'm using Prophesee Gen3 camera now, and having a feature request to you. Since I couldn't find any forum or feature request form for Metavision SDK, I'm writing here.

What I want to do in the end: load .raw data on my Mac (M1, 2020). And currently this is not possible in my understanding.

What I want to request: either of the below options:

• Please support Mac for Metavision SDK. Then I can install the python library.
• Please develop a python-native library (something like pip install metavision_api). I don't need many functionalities for now, just data loading is a great help. I believe this is possible without Metavision backend.

Sincerely,

Shintaro

Hi,

I'm having problems with the compiling tests and I have uploaded the command line output I got.
test_log.txt

The first test passed but the second was showing AssignProcessToJobObject: (87) The parameter is incorrect., and the error stopped the whole testing process.

There was no error while building the code. I'm on Windows 10 with Visual Studio 2022, and I used standalone CMake 3.22.1. I have also installed the HAL plugin as I'm using a Prohphesee camera.

I was using the Essentials SDK before but I had some problems with Release/Debug configurations, and it was uninstalled before I started with OpenEB.

Thank you very much, Phil

Hi, I was looking into buying a SilkyEVCam to be used on a robot that runs an Nvidia Xavier AGX (ARM64 based system). Is that supported by OpenEB? Will it be possible to get a HAL plugin for ARM64/Ubuntu 18.04 systems? Thanks!

Hey!

Happy to see you opened source code of plugins. For a long time, to install plugins in different CPU architectures platforms are annoyed me. I'm trying use openeb in my robot project. But I found the installation scripts are not allowd me to install this without root privilege because of the USB rules installation hal_psee_plugins/resources/rules/CMakeLists.txt.

AFAIK. The rules are usually installed independently with shell scripts, like setup_udev_rules.sh in librealsense project. So perhaps create a new file for rules installation is a better idea?

Hi, thank you for your support. I have two questions about python in compilation.

Environment

lsb_release -a
No LSB modules are available.
Distributor ID:	Ubuntu
Description:	Ubuntu 20.04.3 LTS
Release:	20.04
Codename:	focal

python

I have python3.8 and 3.9:

➜  build git:(main) ✗ which python
python not found
➜  build git:(main) ✗ which python3
/usr/bin/python3
➜  build git:(main) ✗ which python3.8
/usr/bin/python3.8
➜  build git:(main) ✗ which python3.9
/usr/bin/python3.9
➜  build git:(main) ✗ python3 -V
Python 3.8.10

Q1. Compilation fails because it does not find python3.8

I think cmake tries to find python3.9 first although I have 3.8. How can I tell the pythonpath and/or python version?

Here is the error:

build git:(main) ✗ cmake .. -DBUILD_TESTING=OFF
-- The C compiler identification is GNU 9.3.0
-- The CXX compiler identification is GNU 9.3.0
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Building METAVISION_SDK_MODULES_OPEN modules : base;core;driver;ui
-- Building METAVISION_SDK_MODULES_PROFESSIONAL modules : analytics;calibration;cv;ml;3dview;cv3d
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed
-- Check if compiler accepts -pthread
-- Check if compiler accepts -pthread - yes
-- Found Threads: TRUE  
-- Found PkgConfig: /usr/bin/pkg-config (found version "0.29.1") 
-- Checking for module 'libusb-1.0'
--   Found libusb-1.0, version 1.0.23
-- Found LibUSB: /usr/include/libusb-1.0  
-- Found Boost: /usr/lib/x86_64-linux-gnu/cmake/Boost-1.71.0/BoostConfig.cmake (found version "1.71.0") found components: program_options filesystem timer chrono thread 
CMake Error at /usr/share/cmake-3.16/Modules/FindPackageHandleStandardArgs.cmake:146 (message):
  Could NOT find Python3 (missing: Python3_INCLUDE_DIRS Python3_LIBRARIES
  Development Development.Module Development.Embed) (found version "3.9.7")
Call Stack (most recent call first):
  /usr/share/cmake-3.16/Modules/FindPackageHandleStandardArgs.cmake:393 (_FPHSA_FAILURE_MESSAGE)
  cmake/Modules/FindPython/Support.cmake:2999 (find_package_handle_standard_args)
  cmake/Modules/FindPython3.cmake:391 (include)
  cmake/custom_functions/python3.cmake:63 (find_package)
  CMakeLists.txt:211 (include)


-- Configuring incomplete, errors occurred!
See also "/home/shintaro/prophesee_ws/openeb/build/CMakeFiles/CMakeOutput.log".
See also "/home/shintaro/prophesee_ws/openeb/build/CMakeFiles/CMakeError.log".

I tried this issue but didn't work either. https://github.com/prophesee-ai/openeb/issues/32#issuecomment-945825601

Q2. virtual environment

My existing python environment is already coupled with ROS and some DAVIS packages. Since your library requirement is somehow strict ("numpy==1.19.5" "opencv-python>=4.2.0.34"), I'd like to create virtual environment to encapsulate openeb environment. I'm thinking something like python3.8 -m venv ~/local/venv but poetry is also fine.

Have you every tried it? How can I make it work with venv?

Sincerely,

Shintaro

Hi. I recently recorded a series of event data under ROS using the prophesee_ros_wrapper. This choice is because ROS provides a versatile framework that supports multi-sensors recording. However, I need to transform events from ROS bags to HDF5 files for now, as the ROS bag is quite bulky. Is there any function implemented in the metavision, or sample script to do so?

Thanks in advance!

I want to run Blinking Lights detector samples in local PyCharm IDE. The OpenEB library has been installed and compiled. However, I can not find BaseFrameGenerationAlgorithm in the metavision_sdk_core module.

Could you help me to solve this problem?

Hi there, I think I found a bug in the metavision API:

Environment:

• Ubuntu 20.04

------------------------------------------ 
PLATFORM INFORMATION 
------------------------------------------ 


####  DISTRIBUTION  #### 

Linux scioi244-34 5.15.0-48-generic #54~20.04.1-Ubuntu SMP Thu Sep 1 16:17:26 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux 

####  VIRTUAL MACHINE  #### 

Virtual Machine:              NO
Docker:                       NO

------------------------------------------ 
METAVISION SOFTWARE INFORMATION 
------------------------------------------ 


####  INSTALLED SOFTWARE  #### 

Version:                      3.0.2
Id:                           f3107c5b0cfacb2ab97aa78ebefcdc3b4992c7c0

####  INSTALLED PACKAGES - METAVISION  #### 

Failed to retrieve installed Metavision packages list 

------------------------------------------ 
METAVISION SYSTEMS INFORMATION 
------------------------------------------ 


####  SYSTEMS AVAILABLE  #### 

No systems USB connected have been found on your platform. 

Bug

We found an inconsistency when using RawReader and the seek_event method. Here is a minimal example. This first example works fine:

import numpy as np
from metavision_core.event_io import EventsIterator, RawReader

raw_path = "/tmp/spinner.raw"
iterator = EventsIterator(raw_path, mode="delta_t")
reader = RawReader(raw_path)

for i, events_gt in enumerate(iterator):
    n_events = len(events_gt)
    events = reader.load_n_events(n_events)

    print(f"batch: {i}\niterator: {events_gt[:3]}\nraw_reader: {events[:3]}\n")

    events_gt = np.stack([events_gt["y"], events_gt["x"], events_gt["t"], events_gt["p"]]).T
    events = np.stack([events["y"], events["x"], events["t"], events["p"]]).T
    assert np.allclose(events_gt, events)

producing the output:

batch: 0
iterator: [(237, 121, 1, 0) (246, 121, 1, 0) (248, 132, 1, 0)]
raw_reader: [(237, 121, 1, 0) (246, 121, 1, 0) (248, 132, 1, 0)]

batch: 1
iterator: [(377, 98, 1, 10000) (359, 98, 1, 10000) (349, 97, 1, 10000)]
raw_reader: [(377, 98, 1, 10000) (359, 98, 1, 10000) (349, 97, 1, 10000)]

batch: 3
iterator: [(332, 309, 1, 30000) (357, 310, 1, 30000) (358, 310, 1, 30000)]
raw_reader: [(332, 309, 1, 30000) (357, 310, 1, 30000) (358, 310, 1, 30000)]

batch: 4
iterator: [(249, 287, 1, 40000) (204, 237, 0, 40000) (219, 237, 0, 40000)]
raw_reader: [(249, 287, 1, 40000) (204, 237, 0, 40000) (219, 237, 0, 40000)]

...

However, this snippet produces two bugs:

import numpy as np
from metavision_core.event_io import EventsIterator, RawReader

raw_path = "/tmp/spinner.raw"
iterator = EventsIterator(raw_path, mode="delta_t")
reader = RawReader(raw_path)
cnt = 0

for i, events_gt in enumerate(iterator):
    n_events = len(events_gt)

    reader.reset()
    reader.seek_event(cnt)

    events = reader.load_n_events(n_events)
    cnt += n_events

    print(f"batch: {i}\niterator: {events_gt[:3]}\nraw_reader: {events[:3]}\n")

    events_gt = np.stack([events_gt["y"], events_gt["x"], events_gt["t"], events_gt["p"]]).T
    events = np.stack([events["y"], events["x"], events["t"], events["p"]]).T

    if i:
        assert np.allclose(events_gt, events)

the output is:

batch: 0
iterator: [(237, 121, 1, 0) (246, 121, 1, 0) (248, 132, 1, 0)]
raw_reader: [(  0,   0, 0, 0) (237, 121, 1, 0) (246, 121, 1, 0)]

batch: 1
iterator: [(377, 98, 1, 10000) (359, 98, 1, 10000) (349, 97, 1, 10000)]
raw_reader: [(377, 98, 1, 10000) (359, 98, 1, 10000) (349, 97, 1, 10000)]

batch: 2
iterator: [(432, 228, 1, 20000) (426, 230, 1, 20000) (432, 231, 1, 20000)]
raw_reader: [(432, 228, 1, 20000) (426, 230, 1, 20000) (432, 231, 1, 20000)]

batch: 3
iterator: [(332, 309, 1, 30000) (357, 310, 1, 30000) (358, 310, 1, 30000)]
raw_reader: [(332, 309, 1, 30000) (357, 310, 1, 30000) (358, 310, 1, 30000)]

batch: 4
iterator: [(249, 287, 1, 40000) (204, 237, 0, 40000) (219, 237, 0, 40000)]
raw_reader: [(249, 287, 1, 40000) (204, 237, 0, 40000) (219, 237, 0, 40000)]

batch: 5
iterator: [(210, 142, 1, 50000) (220, 142, 1, 50000) (227, 142, 1, 50000)]
raw_reader: [(210, 142, 1, 50000) (220, 142, 1, 50000) (227, 142, 1, 50000)]

batch: 6
iterator: [(342, 110, 1, 60000) (376, 110, 0, 60000) (343, 111, 1, 60000)]
raw_reader: [(342, 110, 1, 60000) (376, 110, 0, 60000) (343, 111, 1, 60000)]

batch: 7
iterator: [(417, 187, 1, 70000) (423, 187, 1, 70000) (420, 186, 1, 70000)]
raw_reader: [(417, 187, 1, 70000) (423, 187, 1, 70000) (420, 186, 1, 70000)]

batch: 8
iterator: [(358, 293, 1, 80000) (361, 292, 1, 80000) (352, 292, 1, 80000)]
raw_reader: [(358, 293, 1, 80000) (361, 292, 1, 80000) (352, 292, 1, 80000)]

batch: 9
iterator: [(210, 256, 0, 90000) (240, 278, 1, 90000) (246, 278, 1, 90000)]
raw_reader: [(210, 256, 0, 90000) (240, 278, 1, 90000) (246, 278, 1, 90000)]

batch: 10
iterator: [(207, 141, 1, 100000) (239, 140, 0, 100000) (224, 160, 1, 100000)]
raw_reader: [(207, 141, 1, 100000) (239, 140, 0, 100000) (224, 160, 1, 100000)]

batch: 11
iterator: [(331, 98, 1, 110000) (322, 98, 1, 110000) (359, 98, 0, 110000)]
raw_reader: [(331, 98, 1, 110000) (322, 98, 1, 110000) (359, 98, 0, 110000)]

batch: 12
iterator: [(430, 184, 1, 120000) (429, 184, 1, 120000) (410, 188, 1, 120000)]
raw_reader: [(430, 184, 1, 120000) (429, 184, 1, 120000) (410, 188, 1, 120000)]

batch: 13
iterator: [(376, 300, 1, 130000) (376, 302, 1, 130000) (343, 302, 0, 130000)]
raw_reader: [(376, 300, 1, 130000) (376, 302, 1, 130000) (343, 302, 0, 130000)]

batch: 14
iterator: [(222, 271, 0, 140000) (224, 269, 0, 140000) (232, 268, 0, 140000)]
raw_reader: [(222, 271, 0, 140000) (224, 269, 0, 140000) (232, 268, 0, 140000)]

batch: 15
iterator: [(228, 146, 0, 150000) (223, 134, 0, 150000) (234, 141, 0, 150000)]
raw_reader: [(209, 169, 0, 250027) (213, 169, 0, 250027) (215, 169, 0, 250027)]

Traceback (most recent call last):
  File "minimal-error.py", line 24, in <module>
    assert np.allclose(events_gt, events)
AssertionError

The first problem is the (0, 0, 0, 0) event inserted in the first batch. The second problem is a jump in the event data in batch 15.

Hello,

I followed the instructions in this repository to install OpenEB on my Ubuntu 18.04. But I got errors after I did cmake --build . --config Release -- -j 4. Here is a part of the summary:

/home/kamzero/Downloads/openeb/sdk/modules/core/cpp/src/video_writer.cpp: In member function ‘virtual void Metavision::VideoWriter::write(cv::InputArray)’: /home/kamzero/Downloads/openeb/sdk/modules/core/cpp/src/video_writer.cpp:343:33: error: no matching function for call to ‘Metavision::VideoWriter::write(const cv::_InputArray&)’ cv::VideoWriter::write(image); ^ In file included from /home/kamzero/Downloads/openeb/sdk/modules/core/cpp/src/video_writer.cpp:51:0: /usr/local/include/opencv2/videoio.hpp:940:26: note: candidate: virtual void cv::VideoWriter::write(const cv::Mat&) CV_WRAP virtual void write(const Mat& image); ^~~~~ /usr/local/include/opencv2/videoio.hpp:940:26: note: no known conversion for argument 1 from ‘const cv::_InputArray’ to ‘const cv::Mat&’ /home/kamzero/Downloads/openeb/sdk/modules/core/cpp/src/video_writer.cpp: In member function ‘virtual Metavision::VideoWriter& Metavision::VideoWriter::operator<<(const cv::UMat&)’: /home/kamzero/Downloads/openeb/sdk/modules/core/cpp/src/video_writer.cpp:365:72: error: no matching function for call to ‘Metavision::VideoWriter::operator<<(const cv::UMat&)’ return static_cast<VideoWriter &>(cv::VideoWriter::operator<<(image)); ^ In file included from /home/kamzero/Downloads/openeb/sdk/modules/core/cpp/src/video_writer.cpp:51:0: /usr/local/include/opencv2/videoio.hpp:931:26: note: candidate: virtual cv::VideoWriter& cv::VideoWriter::operator<<(const cv::Mat&) virtual VideoWriter& operator << (const Mat& image); ^~~~~~~~ /usr/local/include/opencv2/videoio.hpp:931:26: note: no known conversion for argument 1 from ‘const cv::UMat’ to ‘const cv::Mat&’ sdk/modules/core/cpp/lib/CMakeFiles/metavision_sdk_core.dir/build.make:185: recipe for target 'sdk/modules/core/cpp/lib/CMakeFiles/metavision_sdk_core.dir/__/src/video_writer.cpp.o' failed make[2]: *** [sdk/modules/core/cpp/lib/CMakeFiles/metavision_sdk_core.dir/__/src/video_writer.cpp.o] Error 1 CMakeFiles/Makefile2:2203: recipe for target 'sdk/modules/core/cpp/lib/CMakeFiles/metavision_sdk_core.dir/all' failed make[1]: *** [sdk/modules/core/cpp/lib/CMakeFiles/metavision_sdk_core.dir/all] Error 2 make[1]: *** Waiting for unfinished jobs.... [ 35%] Built target metavision_sdk_base_python3 [ 36%] Linking CXX executable ../../../../bin/metavision_raw_cutter [ 36%] Built target metavision_raw_cutter [ 37%] Linking CXX executable ../../../../bin/metavision_platform_info [ 37%] Built target metavision_platform_info Makefile:170: recipe for target 'all' failed make: *** [all] Error 2

It would be greatly appreciated if you can help.

Hi,

Thank you for releasing this framework. I tried to compile and install openeb in my Ubuntu18.04 pc.I followed the instructions in this repository. Compilation goes fine but when I do ctest --verbose I get many errors. The summary is:

38% tests passed, 13 tests failed out of 21

Total Test time (real) =   9.39 sec

The following tests FAILED:
	  1 - metavision-raw-cutter-test (Failed)
	  4 - metavision-raw-info-with-sample-plugin-test (Failed)
	  5 - metavision-raw-cutter-with-sample-plugin-test (Failed)
	  6 - metavision-raw-to-csv-with-sample-plugin-test (Failed)
	  7 - metavision-raw-to-dat-with-sample-plugin-test (Failed)
	 11 - pytests_base (Failed)
	 13 - metavision-raw-to-video-test (Failed)
	 15 - pytests_core (Failed)
	 16 - metavision-raw-info-test (Failed)
	 17 - metavision-raw-to-dat-test (Failed)
	 18 - metavision-raw-to-csv-test (Failed)
	 20 - metavision-evt2-raw-file-decoder-test (Failed)
	 21 - metavision-evt3-raw-file-decoder-test (Failed)

Here is the full output of the ctest command: test.log

I will appreciate if you could guide me on how to fix this errors.

Hi, I'm trying to run metavision_raw_to_csv.py, and compiled openeb as instructed, and ran the setup_env.sh. When I ran metavision_raw_to_csv.py, I got:

Traceback (most recent call last):
  File "metavision_raw_to_csv.py", line 15, in <module>
    from metavision_core.event_io import EventsIterator
  File "/openeb/sdk/modules/core/python/pypkg/metavision_core/event_io/__init__.py", line 12, in <module>
    from .raw_reader import RawReader
  File "/openeb/sdk/modules/core/python/pypkg/metavision_core/event_io/raw_reader.py", line 22, in <module>
    from metavision_hal import DeviceDiscovery, RawFileConfig
ModuleNotFoundError: No module named 'metavision_hal'

Any ideas on how can I solve it? Thanks!

Feature Request!

My team has been successful working with the EVK 2 HD on an Ubuntu desktop. However, our end goal requires us to use a low-power, embedded platform that cannot run Ubuntu.

Are there any plans to support embedded flows like Yocto or Buildroot for OpenEB?