点云realsense

### 使用 Realsense 获取并处理点云

为了利用 Intel RealSense 深度相机（如D435i型号）来获取和保存点云数据，可以采用 librealsense 库中的 `rs2::pointcloud` 类创建点云对象，并通过调用相应的方法计算实际的点云数据[^1]。

下面是一个简单的 C++ 示例程序用于展示如何从 RealSense 设备读取深度帧并转换成点云：

#include <librealsense2/rs.hpp>    // Include RealSense Cross Platform API Headers
#include <iostream>
#include "pcl/visualization/cloud_viewer.h"

int main() try {
    rs2::pipeline pipe;
    auto config = rs2::config();
    
    // Configure streams to enable
    config.enable_stream(RS2_STREAM_DEPTH, 640, 480, RS2_FORMAT_Z16, 30);
    pipe.start(config);

    using namespace std;

    cout << "Press ESC key to stop capturing..." << endl;

    pcl::PointCloud<pcl::PointXYZ>::Ptr cloud(new pcl::PointCloud<pcl::PointXYZ>);
    rs2::pointcloud pc; // Create pointcloud object
    while (true) {

        // Wait for next set of frames from the camera
        rs2::frameset data = pipe.wait_for_frames();

        // Generate the pointcloud and texture mappings
        rs2::points points = pc.calculate(data.get_depth_frame());

        // Get vertices as vector of float triples
        const auto verts = points.get_vertices();

        // Clear previous content before adding new points.
        cloud->clear();
        
        // Add each vertex into PCL PointCloud structure
        for(auto& vert : verts){
            if (!std::isnan(vert.x)){
                cloud->push_back(pcl::PointXYZ(vert.x, vert.y, vert.z));
            }
        }

        // Visualization with PCL viewer
        pcl::visualization::CloudViewer viewer("RealSense Point Cloud Viewer");
        viewer.showCloud(cloud);
        int k = cv::waitKey(1); // Use OpenCV's waitKey function here just for keyboard input handling.

        // Break loop on escape key press
        if(k == 27)
            break;
    }
}
catch (const rs2::error & e){
    std::cerr << "RealSense error calling " << e.get_failed_function() << "(" << e.get_failed_args() << "):\n    " << e.what() << std::endl;
    return EXIT_FAILURE;
}
catch (const std::exception& e){
    std::cerr << e.what() << std::endl;
    return EXIT_FAILURE;
}

上述代码片段展示了怎样设置管道以接收来自 RealSense 的深度流，并使用 `rs2::pointcloud` 对象将这些深度信息转化为三维空间坐标系下的点集。最后借助 PCL 库实现了基本可视化功能。