run build.sh failed - Githubissues

PIGzyj233 commented 5 years ago

CMakeFiles/Makefile2:72: recipe for target 'CMakeFiles/ORB_SLAM2.dir/all' failed make[1]: [CMakeFiles/ORB_SLAM2.dir/all] Error 2 Makefile:83: recipe for target 'all' failed make: [all] Error 2 Converting vocabulary to binary BoW load/save benchmark ./tools/bin_vocabulary: symbol lookup error: ./tools/bin_vocabulary: undefined symbol: _ZN5DBoW24FORB10fromStringERN2cv3MatERKSs

environment: ubuntu16.04 cuda10.1 opencv2.4.11 pcl1.8.0 CMakeError.log CMakeOutput.log

IoannisAsmanis commented 5 years ago

Fairly certain I've ran into the same exact issue...

CMakeFiles/ORB_SLAM2.dir/build.make:542: recipe for target 'CMakeFiles/ORB_SLAM2.dir/src/segmentation.cc.o' failed
make[2]: *** [CMakeFiles/ORB_SLAM2.dir/src/segmentation.cc.o] Error 1
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/ORB_SLAM2.dir/all' failed
make[1]: *** [CMakeFiles/ORB_SLAM2.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2
Converting vocabulary to binary
BoW load/save benchmark
./tools/bin_vocabulary: symbol lookup error: ./tools/bin_vocabulary: undefined symbol: _ZN5DBoW24FORB10fromStringERN2cv3MatERKSs

Ubuntu 16.04, CUDA 10.1, pcl1.8.1, opencv2.4.9

Any luck in overcoming this yet?

qixuxiang commented 5 years ago

hi @PIGzyj233 @IoannisAsmanis The system was not tested on CUDA10, you can try it on CUDA8/9.

PIGzyj233 commented 5 years ago

hi @PIGzyj233 @IoannisAsmanis The system was not tested on CUDA10, you can try it on CUDA8/9. thank you for your reply I successfully compiled the program with cuda9.0, but there is no pcl viewer display, and soon the same as the previous user reported "segment error, the core has been dumped" after tracking, the display is libopencv.so.2.4.13 and libc.so .6 has an error Can you tell me what environment configuration you are running successfully? Such as cmake version, pcl version and opencv version

qixuxiang commented 5 years ago

@PIGzyj233 PCL 1.8.1，CUDA 8 and opencv 2.3.13 On Ubuntu 16.04

Vslamer commented 5 years ago

遇到了同样的问题，CUDA 10.1 Ubuntu 18.04 是cuda版本的问题吗

qixuxiang commented 5 years ago

暂未测试cuda10，建议回退到cuda8/cuda9 @Moonkisscy

Vslamer commented 5 years ago

`make[2]: 没有规则可制作目标“../Thirdparty/darknet/build/libYOLOv3SE.so”，由“../lib/libORB_SLAM2.so” 需求。停止。 CMakeFiles/Makefile2:104: recipe for target 'CMakeFiles/ORB_SLAM2.dir/all' failed make[1]: [CMakeFiles/ORB_SLAM2.dir/all] Error 2 Makefile:83: recipe for target 'all' failed make: *** [all] Error 2 Converting vocabulary to binar build.sh: 46: build.sh: ./tools/bin_vocabulary: not found

` 我的报错是这个，退为cuda9，挺麻烦的，想问下有其他办法或者是其他问题，修改代码之类。

qixuxiang commented 5 years ago

CMakeLists.txt里面最后注释全部去掉：

# Build tools
#set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${PROJECT_SOURCE_DIR}/tools)
#add_executable(bin_vocabulary
#tools/bin_vocabulary.cc)
#target_link_libraries(bin_vocabulary ${PROJECT_NAME})

改为：

# Build tools
set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${PROJECT_SOURCE_DIR}/tools)
add_executable(bin_vocabulary
tools/bin_vocabulary.cc)
target_link_libraries(bin_vocabulary ${PROJECT_NAME})

Vslamer commented 5 years ago

好的，谢谢。刚改了，但还是有问题，把所有报错贴在了下面，看上面说没找到库libOpenNI2，但是查看这个库是有的。

CMake Warning at CMakeLists.txt:350 (add_executable):
  Cannot generate a safe linker search path for target bin_vocabulary because
  files in some directories may conflict with libraries in implicit
  directories:

    link library [libOpenNI2.so] in /usr/lib may be hidden by files in:
      /home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/OpenNI-Linux-x64-2.3/Redist

  Some of these libraries may not be found correctly.

CMake Warning at CMakeLists.txt:312 (add_executable):
  Cannot generate a safe linker search path for target rgbd_tum because files
  in some directories may conflict with libraries in implicit directories:

    link library [libOpenNI2.so] in /usr/lib may be hidden by files in:
      /home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/OpenNI-Linux-x64-2.3/Redist

  Some of these libraries may not be found correctly.

CMake Warning at CMakeLists.txt:135 (add_library):
  Cannot generate a safe linker search path for target ORB_SLAM2 because
  files in some directories may conflict with libraries in implicit
  directories:

    link library [libOpenNI2.so] in /usr/lib may be hidden by files in:
      /home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/OpenNI-Linux-x64-2.3/Redist

  Some of these libraries may not be found correctly.

-- Generating done
-- Build files have been written to: /home/cy/dlslam/orb-slam2_with_semantic_label/build
make[2]: *** 没有规则可制作目标“../Thirdparty/darknet/build/libYOLOv3SE.so”，由“../lib/libORB_SLAM2.so” 需求。 停止。
CMakeFiles/Makefile2:141: recipe for target 'CMakeFiles/ORB_SLAM2.dir/all' failed
make[1]: *** [CMakeFiles/ORB_SLAM2.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2
Converting vocabulary to binary
build.sh: 46: build.sh: ./tools/bin_vocabulary: not found

qixuxiang commented 5 years ago

@Moonkisscy [libOpenNI2.so]这个可以注释掉，这个库可选，并非必须

qixuxiang commented 5 years ago

@Moonkisscy 在原始Cmakelist.txt 第280行

#${OPENNI_LIB_DIR}/libOpenNI2.so

Vslamer commented 5 years ago

@Moonkisscy [libOpenNI2.so]这个可以注释掉，这个库可选，并非必须

嗯，刚注释了，库没报错了，现在还是之前这个问题。

make[2]: *** 没有规则可制作目标“../Thirdparty/darknet/build/libYOLOv3SE.so”，由“../lib/libORB_SLAM2.so” 需求。 停止。
CMakeFiles/Makefile2:141: recipe for target 'CMakeFiles/ORB_SLAM2.dir/all' failed
make[1]: *** [CMakeFiles/ORB_SLAM2.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2
Converting vocabulary to binary
build.sh: 46: build.sh: ./tools/bin_vocabulary: not found

qixuxiang commented 5 years ago

@Moonkisscy 算了，饶了弯路，直接在build.sh删掉

echo "Converting vocabulary to binary"
./tools/bin_vocabulary`

这两行。然后重新在CMakeLists.txt最后加上注释

# Build tools
#set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${PROJECT_SOURCE_DIR}/tools)
#add_executable(bin_vocabulary
#tools/bin_vocabulary.cc)
#target_link_libraries(bin_vocabulary ${PROJECT_NAME})

最后进入Vocabulary执行：tar -xzvf ORBvoc.txt.tar.gz 直接进入Readme第三步run code

如果可以跑成功，记得回来回复下。

Vslamer commented 5 years ago

试了下，还是有问题额，难道真是cuda版本问题吗。

-- Build files have been written to: /home/cy/dlslam/orb-slam2_with_semantic_label/build
make[2]: *** 没有规则可制作目标“../Thirdparty/darknet/build/libYOLOv3SE.so”，由“../lib/libORB_SLAM2.so” 需求。 停止。
CMakeFiles/Makefile2:104: recipe for target 'CMakeFiles/ORB_SLAM2.dir/all' failed
make[1]: *** [CMakeFiles/ORB_SLAM2.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2

qixuxiang commented 5 years ago

@Moonkisscy 这里需要把libYOLOv3SE.so编译出来，手动执行

cd Thirdparty/darknet
mkdir build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
make -j2

Vslamer commented 5 years ago

@Moonkisscy 这里需要把libYOLOv3SE.so编译出来，手动执行
cd Thirdparty/darknet
mkdir build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
make -j2

按照上说的编译又遇到下面错误，查了下说是是因为缺乏头文件cudnn.h，需要将其复制。于是运行：`sudo cp cudnn.h /usr/include复制。又遇到“cp: 无法获取'cudnn.h' 的文件状态(stat): 没有那个文件或目录”错误。不知道是不是这样解决，因为刚接触这块不久，所以问题可能有点多，望见谅，哈哈。

In file included from /home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/darknet/src/layer.h:1:0,
                 from /home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/darknet/src/dropout_layer.h:4,
                 from /home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/darknet/src/dropout_layer_kernels.cu:6:
/home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/darknet/src/darknet.h:46:23: fatal error: cudnn.h: 没有那个文件或目录
compilation terminated.
In file included from /home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/darknet/src/layer.h:1:0,
                 from /home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/darknet/src/dropout_layer.h:4,
                 from /home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/darknet/src/dropout_layer_kernels.cu:6:
/home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/darknet/src/darknet.h:46:23: fatal error: cudnn.h: 没有那个文件或目录
compilation terminated.
CMake Error at YOLOv3SE_Train_generated_dropout_layer_kernels.cu.o.Release.cmake:219 (message):
  Error generating
  /home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/darknet/build/CMakeFiles/YOLOv3SE_Train.dir/src/./YOLOv3SE_Train_generated_dropout_layer_kernels.cu.o

CMake Error at YOLOv3SE_generated_dropout_layer_kernels.cu.o.Release.cmake:219 (message):
  Error generating
  /home/cy/dlslam/orb-slam2_with_semantic_label/Thirdparty/darknet/build/CMakeFiles/YOLOv3SE.dir/src/./YOLOv3SE_generated_dropout_layer_kernels.cu.o

CMakeFiles/YOLOv3SE.dir/build.make:196: recipe for target 'CMakeFiles/YOLOv3SE.dir/src/YOLOv3SE_generated_dropout_layer_kernels.cu.o' failed
make[2]: *** [CMakeFiles/YOLOv3SE.dir/src/YOLOv3SE_generated_dropout_layer_kernels.cu.o] Error 1
CMakeFiles/YOLOv3SE_Train.dir/build.make:126: recipe for target 'CMakeFiles/YOLOv3SE_Train.dir/src/YOLOv3SE_Train_generated_dropout_layer_kernels.cu.o' failed
make[2]: *** [CMakeFiles/YOLOv3SE_Train.dir/src/YOLOv3SE_Train_generated_dropout_layer_kernels.cu.o] Error 1
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/YOLOv3SE.dir/all' failed
make[1]: *** [CMakeFiles/YOLOv3SE.dir/all] Error 2
make[1]: *** 正在等待未完成的任务....
CMakeFiles/Makefile2:104: recipe for target 'CMakeFiles/YOLOv3SE_Train.dir/all' failed
make[1]: *** [CMakeFiles/YOLOv3SE_Train.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2

qixuxiang commented 5 years ago

@Moonkisscy

fatal error: cudnn.h: 没有那个文件或目录

这个问题尝试自己解决下吧，解决完就可以跑起来了。

欢迎入坑SLAM

Vslamer commented 5 years ago

@Moonkisscy
fatal error: cudnn.h: 没有那个文件或目录
这个问题尝试自己解决下吧，解决完就可以跑起来了。

欢迎入坑SLAM

已入坑，哈哈。这个问题解决了，又遇到一个新问题，谷歌了挺久没找到合适方法解决，不知道你有没有遇到过。

make[2]: *** 没有规则可制作目标“/usr/local/cuda/lib64/libcudnn.so”，由“libYOLOv3SE.so” 需求。 停止。
make[2]: *** 正在等待未完成的任务....
[ 97%] Building C object CMakeFiles/YOLOv3SE.dir/src/yolo_v3.c.o
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/YOLOv3SE.dir/all' failed
make[1]: *** [CMakeFiles/YOLOv3SE.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2

qixuxiang commented 5 years ago

搜索下安装linux下安装cudnn

Vslamer commented 5 years ago

搜索下安装linux下安装cudnn

恩，这个刚才解决了，又正在解决下面这个问题,感觉解决一个又来一个新的@_@。

/usr/bin/ld: 找不到 -lCUDA_cublas_device_LIBRARY-NOTFOUND
collect2: error: ld returned 1 exit status
CMakeFiles/YOLOv3SE_Train.dir/build.make:3001: recipe for target 'YOLOv3SE_Train' failed
make[2]: *** [YOLOv3SE_Train] Error 1
CMakeFiles/Makefile2:104: recipe for target 'CMakeFiles/YOLOv3SE_Train.dir/all' failed
make[1]: *** [CMakeFiles/YOLOv3SE_Train.dir/all] Error 2

qixuxiang commented 5 years ago

大哥，一个一个来。。。慢慢来，比较快

Vslamer commented 5 years ago

大哥，一个一个来。。。慢慢来，比较快

恩，上面的问题解决了，libYOLOv3SE.so编译没问题了，最后编译ORBSLAM2还是出现下面pcl问题，安装的pcl1.8测试了也没问题。按照网上试了挺多方法，改了CMakeLists.txt等也都出现新问题没有解决，感谢大佬无私回复，哈哈。

[100%] Linking CXX executable ../bin/rgbd_tum
../lib/libORB_SLAM2.so：对‘pcl::search::Search<pcl::PointXYZ>::getName() const’未定义的引用
../lib/libORB_SLAM2.so：对‘pcl::search::Search<pcl::PointXYZRGBNormal>::getName() const’未定义的引用
../lib/libORB_SLAM2.so：对‘pcl::search::Search<pcl::PointXYZRGB>::getName() const’未定义的引用
../lib/libORB_SLAM2.so：对‘pcl::search::Search<pcl::PointXYZRGBA>::getName() const’未定义的引用
../lib/libORB_SLAM2.so：对‘pcl::search::Search<pcl::PointXYZI>::getName() const’未定义的引用
collect2: error: ld returned 1 exit status
CMakeFiles/rgbd_tum.dir/build.make:352: recipe for target '../bin/rgbd_tum' failed
make[2]: *** [../bin/rgbd_tum] Error 1
CMakeFiles/Makefile2:109: recipe for target 'CMakeFiles/rgbd_tum.dir/all' failed
make[1]: *** [CMakeFiles/rgbd_tum.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2
Converting vocabulary to binary
BoW load/save benchmark
Loading fom text: 5.85s
Saving as binary: 0.09s

450932894 commented 5 years ago

[100%] Linking CXX executable ../bin/rgbd_tum ../lib/libORB_SLAM2.so：对‘pcl::SupervoxelClustering::SupervoxelClustering(float, float)’未定义的引用 ../lib/libORB_SLAM2.so：对‘pcl::SupervoxelClustering::setUseSingleCameraTransform(bool)’未定义的引用 collect2: error: ld returned 1 exit status CMakeFiles/rgbd_tum.dir/build.make:403: recipe for target '../bin/rgbd_tum' failed make[2]: [../bin/rgbd_tum] Error 1 CMakeFiles/Makefile2:109: recipe for target 'CMakeFiles/rgbd_tum.dir/all' failed make[1]: [CMakeFiles/rgbd_tum.dir/all] Error 2 Makefile:83: recipe for target 'all' failed make: *** [all] Error 2 Converting vocabulary to binary build.sh: 46: build.sh: ./tools/bin_vocabulary: not found

非常欣赏你的工程，还望解答。

qixuxiang commented 5 years ago

@450932894 请附加你的软硬件信息

450932894 commented 5 years ago

environment: CMakeLists.txt

ubuntu16.04 cuda9.0 opencv3.4.1 pcl1.8.1

450932894 commented 5 years ago

这个问题解决了，但是碰到了新的问题 tang@tangdingxin:~/catkin_ws/src/orb-slam2_with_semantic_label/bin$ ./rgbd_tum ../Vocabulary/ORBvoc.txt ../Examples/RGB-D/TUM1.yaml ../data/rgbd_dataset_freiburg1_xyz ../data/rgbd_dataset_freiburg1_xyz/associations.txt 段错误 (核心已转储)

450932894 commented 5 years ago

@qixuxiang

qixuxiang commented 5 years ago

@450932894 换tum2或者tum3数据试试，另外保证associations.txt生成正确。

450932894 commented 5 years ago

1 2 3 都试了 associations.txt 文件我用原版的orbslam2跑过

qixuxiang commented 5 years ago

加你微信了

Vslamer commented 5 years ago

这个问题解决了，但是碰到了新的问题 tang@tangdingxin:~/catkin_ws/src/orb-slam2_with_semantic_label/bin$ ./rgbd_tum ../Vocabulary/ORBvoc.txt ../Examples/RGB-D/TUM1.yaml ../data/rgbd_dataset_freiburg1_xyz ../data/rgbd_dataset_freiburg1_xyz/associations.txt 段错误 (核心已转储)

@450932894 我也遇到了这个问题，你后面是怎么解决的呢

-------
Start processing sequence ...
Images in the sequence: 792

New map created with 834 points
receive a keyframe, id = 1
receive a keyframe, id = 2
receive a keyframe, id = 3
receive a keyframe, id = 4
receive a keyframe, id = 5
receive a keyframe, id = 6
receive a keyframe, id = 7
receive a keyframe, id = 8
receive a keyframe, id = 9
receive a keyframe, id = 10
receive a keyframe, id = 11
receive a keyframe, id = 12
receive a keyframe, id = 13
./rgbd_tum: : 未知的错误 -1694371902
OpenCV(3.4.1) Error: Assertion failed (key_ != -1 && "Can't fetch data from terminated TLS container.") in getData, file /home/cy/myslam/opencv/modules/core/src/system.cpp, line 1482
terminate called after throwing an instance of 'cv::Exception'
  what():  OpenCV(3.4.1) /home/cy/myslam/opencv/modules/core/src/system.cpp:1482: error: (-215) key_ != -1 && "Can't fetch data from terminated TLS container." in function getData

已放弃 (核心已转储)

450932894 commented 5 years ago

没解决

发自我的iPhone

------------------ Original ------------------ From: Moonkisscy notifications@github.com Date: Tue,Jul 2,2019 4:35 PM To: qixuxiang/orb-slam2_with_semantic_label orb-slam2_with_semantic_label@noreply.github.com Cc: 450932894 450932894@qq.com, Mention mention@noreply.github.com Subject: Re: [qixuxiang/orb-slam2_with_semantic_label] run build.sh failed (#9)

这个问题解决了，但是碰到了新的问题 tang@tangdingxin:~/catkin_ws/src/orb-slam2_with_semantic_label/bin$ ./rgbd_tum ../Vocabulary/ORBvoc.txt ../Examples/RGB-D/TUM1.yaml ../data/rgbd_dataset_freiburg1_xyz ../data/rgbd_dataset_freiburg1_xyz/associations.txt 段错误 (核心已转储)

@450932894 我也遇到了这个问题，你后面是怎么解决的呢 ------- Start processing sequence ... Images in the sequence: 792 New map created with 834 points receive a keyframe, id = 1 receive a keyframe, id = 2 receive a keyframe, id = 3 receive a keyframe, id = 4 receive a keyframe, id = 5 receive a keyframe, id = 6 receive a keyframe, id = 7 receive a keyframe, id = 8 receive a keyframe, id = 9 receive a keyframe, id = 10 receive a keyframe, id = 11 receive a keyframe, id = 12 receive a keyframe, id = 13 ./rgbdtum: : 未知的错误 -1694371902 OpenCV(3.4.1) Error: Assertion failed (key != -1 && "Can't fetch data from terminated TLS container.") in getData, file /home/cy/myslam/opencv/modules/core/src/system.cpp, line 1482 terminate called after throwing an instance of 'cv::Exception' what(): OpenCV(3.4.1) /home/cy/myslam/opencv/modules/core/src/system.cpp:1482: error: (-215) key_ != -1 && "Can't fetch data from terminated TLS container." in function getData 已放弃 (核心已转储)
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

pengyuan-zhou commented 5 years ago

这个问题解决了，但是碰到了新的问题 tang@tangdingxin:~/catkin_ws/src/orb-slam2_with_semantic_label/bin$ ./rgbd_tum ../Vocabulary/ORBvoc.txt ../Examples/RGB-D/TUM1.yaml ../data/rgbd_dataset_freiburg1_xyz ../data/rgbd_dataset_freiburg1_xyz/associations.txt 段错误 (核心已转储)

@450932894 我也遇到了这个问题，你后面是怎么解决的呢

-------
Start processing sequence ...
Images in the sequence: 792

New map created with 834 points
receive a keyframe, id = 1
receive a keyframe, id = 2
receive a keyframe, id = 3
receive a keyframe, id = 4
receive a keyframe, id = 5
receive a keyframe, id = 6
receive a keyframe, id = 7
receive a keyframe, id = 8
receive a keyframe, id = 9
receive a keyframe, id = 10
receive a keyframe, id = 11
receive a keyframe, id = 12
receive a keyframe, id = 13
./rgbd_tum: : 未知的错误 -1694371902
OpenCV(3.4.1) Error: Assertion failed (key_ != -1 && "Can't fetch data from terminated TLS container.") in getData, file /home/cy/myslam/opencv/modules/core/src/system.cpp, line 1482
terminate called after throwing an instance of 'cv::Exception'
  what():  OpenCV(3.4.1) /home/cy/myslam/opencv/modules/core/src/system.cpp:1482: error: (-215) key_ != -1 && "Can't fetch data from terminated TLS container." in function getData

已放弃 (核心已转储)

请问这个有解决吗，谢谢

linmeeka commented 5 years ago

这个问题解决了，但是碰到了新的问题 tang@tangdingxin:~/catkin_ws/src/orb-slam2_with_semantic_label/bin$ ./rgbd_tum ../Vocabulary/ORBvoc.txt ../Examples/RGB-D/TUM1.yaml ../data/rgbd_dataset_freiburg1_xyz ../data/rgbd_dataset_freiburg1_xyz/associations.txt 段错误 (核心已转储)

@450932894 我也遇到了这个问题，你后面是怎么解决的呢

-------
Start processing sequence ...
Images in the sequence: 792

New map created with 834 points
receive a keyframe, id = 1
receive a keyframe, id = 2
receive a keyframe, id = 3
receive a keyframe, id = 4
receive a keyframe, id = 5
receive a keyframe, id = 6
receive a keyframe, id = 7
receive a keyframe, id = 8
receive a keyframe, id = 9
receive a keyframe, id = 10
receive a keyframe, id = 11
receive a keyframe, id = 12
receive a keyframe, id = 13
./rgbd_tum: : 未知的错误 -1694371902
OpenCV(3.4.1) Error: Assertion failed (key_ != -1 && "Can't fetch data from terminated TLS container.") in getData, file /home/cy/myslam/opencv/modules/core/src/system.cpp, line 1482
terminate called after throwing an instance of 'cv::Exception'
  what():  OpenCV(3.4.1) /home/cy/myslam/opencv/modules/core/src/system.cpp:1482: error: (-215) key_ != -1 && "Can't fetch data from terminated TLS container." in function getData

已放弃 (核心已转储)

请问这个有解决吗，谢谢

请问各位有解决这个问题吗？

qixuxiang commented 5 years ago

@linmeeka @450932894 @zpymyyn you can try opencv2.3.X , rather than opencv 3.4

linmeeka commented 4 years ago

@linmeeka @450932894 @zpymyyn you can try opencv2.3.X , rather than opencv 3.4

Could you tell me the version of FFmpeg you used to compile OpenCV 2.3?

ZhangHengw commented 4 years ago

[100%] Linking CXX executable ../bin/rgbd_tum ../lib/libORB_SLAM2.so：对‘pcl::SupervoxelClusteringpcl::PointXYZRGBA::SupervoxelClustering(float, float)’未定义的引用 ../lib/libORB_SLAM2.so：对‘pcl::SupervoxelClusteringpcl::PointXYZRGBA::setUseSingleCameraTransform(bool)’未定义的引用 collect2: error: ld returned 1 exit status CMakeFiles/rgbd_tum.dir/build.make:403: recipe for target '../bin/rgbd_tum' failed make[2]: [../bin/rgbd_tum] Error 1 CMakeFiles/Makefile2:109: recipe for target 'CMakeFiles/rgbd_tum.dir/all' failed make[1]: [CMakeFiles/rgbd_tum.dir/all] Error 2 Makefile:83: recipe for target 'all' failed make: *** [all] Error 2 Converting vocabulary to binary build.sh: 46: build.sh: ./tools/bin_vocabulary: not found

非常欣赏你的工程，还望解答。

@450932894 你好，pcl链接的那个问题怎么解决的，谢谢

EdwiinJunior commented 4 years ago

Ubuntu16 pcl1.8 opencv2.4 cuda9 cudnn for 9 在主目录的build文件下cmake 后，make时出现错误 orb-slam2_with_semantic_label-master/build$ make make[2]: No rule to make target '/usr/lib/x86_64-linux-gnu/libavdevice.so', needed by '../lib/libORB_SLAM2.so'。停止。 CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/ORB_SLAM2.dir/all' failed make[1]: [CMakeFiles/ORB_SLAM2.dir/all] Error 2 Makefile:83: recipe for target 'all' failed make: *** [all] Error 2

linmeeka commented 4 years ago

[100%] Linking CXX executable ../bin/rgbd_tum ../lib/libORB_SLAM2.so：对‘pcl::SupervoxelClusteringpcl::PointXYZRGBA::SupervoxelClustering(float, float)’未定义的引用 ../lib/libORB_SLAM2.so：对‘pcl::SupervoxelClusteringpcl::PointXYZRGBA::setUseSingleCameraTransform(bool)’未定义的引用 collect2: error: ld returned 1 exit status CMakeFiles/rgbd_tum.dir/build.make:403: recipe for target '../bin/rgbd_tum' failed make[2]: [../bin/rgbd_tum] Error 1 CMakeFiles/Makefile2:109: recipe for target 'CMakeFiles/rgbd_tum.dir/all' failed make[1]: [CMakeFiles/rgbd_tum.dir/all] Error 2 Makefile:83: recipe for target 'all' failed make: *** [all] Error 2 Converting vocabulary to binary build.sh: 46: build.sh: ./tools/bin_vocabulary: not found 非常欣赏你的工程，还望解答。

@450932894 你好，pcl链接的那个问题怎么解决的，谢谢

这个需要带可视化的PCL，先安装VTK再编译PCL。你现在能够成功运行了吗？

dragunov47 commented 4 years ago

[100%] Linking CXX executable ../bin/rgbd_tum ../lib/libORB_SLAM2.so：对‘pcl::SupervoxelClusteringpcl::PointXYZRGBA::SupervoxelClustering(float, float)’未定义的引用 ../lib/libORB_SLAM2.so：对‘pcl::SupervoxelClusteringpcl::PointXYZRGBA::setUseSingleCameraTransform(bool)’未定义的引用 collect2: error: ld returned 1 exit status CMakeFiles/rgbd_tum.dir/build.make:403: recipe for target '../bin/rgbd_tum' failed make[2]: [../bin/rgbd_tum] Error 1 CMakeFiles/Makefile2:109: recipe for target 'CMakeFiles/rgbd_tum.dir/all' failed make[1]: [CMakeFiles/rgbd_tum.dir/all] Error 2 Makefile:83: recipe for target 'all' failed make: *** [all] Error 2 Converting vocabulary to binary build.sh: 46: build.sh: ./tools/bin_vocabulary: not found

非常欣赏你的工程，还望解答。

你好，请问你PCL这个问题怎么解决的？

vayneli commented 4 years ago

这个问题解决了，但是碰到了新的问题 tang@tangdingxin:~/catkin_ws/src/orb-slam2_with_semantic_label/bin$ ./rgbd_tum ../Vocabulary/ORBvoc.txt ../Examples/RGB-D/TUM1.yaml ../data/rgbd_dataset_freiburg1_xyz ../data/rgbd_dataset_freiburg1_xyz/associations.txt 段错误 (核心已转储)

@450932894 我也遇到了这个问题，你后面是怎么解决的呢
-------
Start processing sequence ...
Images in the sequence: 792

New map created with 834 points
receive a keyframe, id = 1
receive a keyframe, id = 2
receive a keyframe, id = 3
receive a keyframe, id = 4
receive a keyframe, id = 5
receive a keyframe, id = 6
receive a keyframe, id = 7
receive a keyframe, id = 8
receive a keyframe, id = 9
receive a keyframe, id = 10
receive a keyframe, id = 11
receive a keyframe, id = 12
receive a keyframe, id = 13
./rgbd_tum: : 未知的错误 -1694371902
OpenCV(3.4.1) Error: Assertion failed (key_ != -1 && "Can't fetch data from terminated TLS container.") in getData, file /home/cy/myslam/opencv/modules/core/src/system.cpp, line 1482
terminate called after throwing an instance of 'cv::Exception'
  what():  OpenCV(3.4.1) /home/cy/myslam/opencv/modules/core/src/system.cpp:1482: error: (-215) key_ != -1 && "Can't fetch data from terminated TLS container." in function getData

已放弃 (核心已转储)
请问这个有解决吗，谢谢
请问各位有解决这个问题吗？请问大佬这个问题解决了吗？，换成opencv2的话，能运行成功吗？

pugna111 commented 4 years ago

environment: CMakeLists.txt

ubuntu16.04 cuda9.0 opencv3.4.1 pcl1.8.1

请问您是怎么解决的？我也碰到这个问题，希望得到你的帮助！谢谢！

vayneli commented 4 years ago

environment: CMakeLists.txt ubuntu16.04 cuda9.0 opencv3.4.1 pcl1.8.1

请问您是怎么解决的？我也碰到这个问题，希望得到你的帮助！谢谢！你的具体报错是什么呢？这个最新的代码应该是可以在18.04下运行成功的，16.04的环境下如果使用libtorch-yolov3的方式替代CMakelists中原来的darknet库应该也能成功，但是具体没测试，建议你可以试试。

bhmujiangyao commented 3 years ago

搜索下安装linux下安装cudnn

恩，这个刚才解决了，又正在解决下面这个问题,感觉解决一个又来一个新的@_@。

/usr/bin/ld: 找不到 -lCUDA_cublas_device_LIBRARY-NOTFOUND
collect2: error: ld returned 1 exit status
CMakeFiles/YOLOv3SE_Train.dir/build.make:3001: recipe for target 'YOLOv3SE_Train' failed
make[2]: *** [YOLOv3SE_Train] Error 1
CMakeFiles/Makefile2:104: recipe for target 'CMakeFiles/YOLOv3SE_Train.dir/all' failed
make[1]: *** [CMakeFiles/YOLOv3SE_Train.dir/all] Error 2

这个是咋解决的嘞？

vayneli commented 3 years ago

建议不要使用run build.sh，cmake，make一步步编译；另外16.04可能有点问题，在18.04下测试可以成功运行。（详见工程的update）

	李佳俊

15008489702@163.com | 签名由网易邮箱大师定制 On 1/15/2021 02:04，bhmujiangyaonotifications@github.com wrote：

搜索下安装linux下安装cudnn

恩，这个刚才解决了，又正在解决下面这个问题,感觉解决一个又来一个新的@_@。

/usr/bin/ld: 找不到 -lCUDA_cublas_device_LIBRARY-NOTFOUND

collect2: error: ld returned 1 exit status

CMakeFiles/YOLOv3SE_Train.dir/build.make:3001: recipe for target 'YOLOv3SE_Train' failed

make[2]: *** [YOLOv3SE_Train] Error 1

CMakeFiles/Makefile2:104: recipe for target 'CMakeFiles/YOLOv3SE_Train.dir/all' failed

make[1]: *** [CMakeFiles/YOLOv3SE_Train.dir/all] Error 2

这个是咋解决的嘞？

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

bhmujiangyao commented 3 years ago

建议不要使用run build.sh，cmake，make一步步编译；另外16.04可能有点问题，在18.04下测试可以成功运行。（详见工程的update） | | 李佳俊 | | 15008489702@163.com | 签名由网易邮箱大师定制 On 1/15/2021 02:04，bhmujiangyaonotifications@github.com wrote：搜索下安装linux下安装cudnn 恩，这个刚才解决了，又正在解决下面这个问题,感觉解决一个又来一个新的@_@。 /usr/bin/ld: 找不到 -lCUDA_cublas_device_LIBRARY-NOTFOUND collect2: error: ld returned 1 exit status CMakeFiles/YOLOv3SE_Train.dir/build.make:3001: recipe for target 'YOLOv3SE_Train' failed make[2]: [YOLOv3SE_Train] Error 1 CMakeFiles/Makefile2:104: recipe for target 'CMakeFiles/YOLOv3SE_Train.dir/all' failed make[1]: [CMakeFiles/YOLOv3SE_Train.dir/all] Error 2 这个是咋解决的嘞？ — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

谢谢 cannot find -lCUDA_cublas_device_LIBRARY-NOTFOUND，就是这个问题不知道咋解决的我是18.04+cuda10.1+libtorch1.4.0,加上libtorch后编译就不成功，不知道是不是版本问题还是啥？

arashi702 commented 3 years ago

2021-07-06 13-46-25屏幕截图请问各位这个问题如何解决 environment:ubuntu18.04+cuda9.0+opencv3.4.5+pcl1.8.1

lin-fangzhou commented 1 year ago

@Moonkisscy
fatal error: cudnn.h: 没有那个文件或目录
这个问题尝试自己解决下吧，解决完就可以跑起来了。欢迎入坑SLAM
已入坑，哈哈。这个问题解决了，又遇到一个新问题，谷歌了挺久没找到合适方法解决，不知道你有没有遇到过。
make[2]: *** 没有规则可制作目标“/usr/local/cuda/lib64/libcudnn.so”，由“libYOLOv3SE.so” 需求。 停止。
make[2]: *** 正在等待未完成的任务....
[ 97%] Building C object CMakeFiles/YOLOv3SE.dir/src/yolo_v3.c.o
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/YOLOv3SE.dir/all' failed
make[1]: *** [CMakeFiles/YOLOv3SE.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2

这个是怎么解决的？能不能分享一下解决方法呢？

dragunov47 commented 1 year ago

您好，我是刘文杰。邮件已收到，谢谢！

lin-fangzhou commented 1 year ago

您好，我是刘文杰。邮件已收到，谢谢！

hi！你们是找到解决方法了吗？可以请教一下吗？

qixuxiang / orb-slam2_with_semantic_label

run build.sh failed #9