Compiling and Running ORB_SLAM2_SSD_Semantic Project
Environment
Ubuntu 20.04 vtk-7.1.1 pcl-1.8.0
Issue 1
Resolution: Modify vtkAtomic.h as follows:
// typedef detail::AtomicOps<sizeof(T*)> Impl;
typedef ::detail::AtomicOps<sizeof(T*)> Impl;
Issue 2
Resolution:
Check for the presence of libvtkFiltersParallelDIY2 in /usr/local/lib. If only libvtkFiltersParallelDIY2-7.1.so exists, create a symbolic link:
sudo ln -sv libvtkFiltersParallelDIY2-7.1.so libvtkFiltersParallelDIY2.so
Repeat for other missing VTK libraries.
Issue 3
Download ncnn:
git clone https://github.com/Tencent/ncnn.git
Navigate to the ncnn directory and compile:
mkdir build
cd build
cmake ..
make -j12
After compilation, copy the library files:
cd ncnn/build/src
sudo cp -n libncnn.a /usr/include
sudo cp -n libncnn.a /usr/lib
Issue 4
Resolution: Obtain camport3:
git clone https://github.com/percipioxyz/camport3.git
Copy shared library files:
cd camport3/lib/linux/lib_x64
sudo cp -n libtycam.so /usr/include
sudo cp -n libtycam.so /usr/lib
sudo cp -n libtycam.so.3.6.12 /usr/include
sudo cp -n libtycam.so.3.6.12 /usr/lib
Issue 5
Resolution: Update CMakeLists.txt in the perfect folder:
add_executable(rgbd_ty
Examples/RGB-D/my_rgbd.cc)
target_link_libraries(rgbd_ty
${PROJECT_NAME}
${TY_TARGET_LIB})
//======= Change to ========
add_executable(rgbd_tum
Examples/RGB-D/rgbd_tum.cc)
target_link_libraries(rgbd_tum
${PROJECT_NAME}
${TY_TARGET_LIB})
Additional ncnn Setup
Run make install after building ncnn. Then, locate the install directory under build and update the header file path in perfect/src/Detector.h to point to the installed ncnn headers.
Compilation should now succeed.