This repository aims to enable the use of Point Cloud Library (PCL) in a zero-copy manner with other cloud stacks. By using a container other than a std::vector, it reduces the overhead of constructors/destructors and resizing.
- Resolved issue with finding CUDA libraries
- Implemented tk_allocator to allow mapping of pre-existing memory
- Fixed compilation errors on ARM architecture
- Fixed issue with finding Google Test framework
pcl::PointCloud<pcl::PointXYZ>::Ptr cloud(new pcl::PointCloud<pcl::PointXYZ>());
std::array<pcl::PointXYZ, 460400> data;
cloud->points = std::vector<pcl::PointXYZ, tk::tk_allocator<pcl::PointXYZ>>(0, tk::tk_allocator<pcl::PointXYZ>(data.data(), 460400));