Pixhawk 开发笔记

drawing

定点降落方案

颜色识别
- 优点：简单、实时性好
- 缺点：容易受光照干扰
形状检测
- 缺点：识别成功率低，运算量大
AprilTag/ArUco
- 优点：运行稳定，实时性好

ArUco姿态估计实践

以下实验采用Aruco Library库。

第一步：标定

至少采集 15 张图。标定所需的图：

aruco_calibration_grid_board_a4

运行指令：

aruco_calibration_fromimages mycalibrationfile.yml pathToDirWithImage -size 0.03

执行完后生成相机的内参文件 aruco_calibration_grid_board_a4.yml

%YAML:1.0
---
image_width: 1280
image_height: 720
camera_matrix: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [ 1.3293642549176379e+03, 0., 6.5434138777781050e+02, 0.,
       1.3408271477588517e+03, 3.8195972804132447e+02, 0., 0., 1. ]
distortion_coefficients: !!opencv-matrix
   rows: 1
   cols: 5
   dt: d
   data: [ -8.8053880180133173e-02, 2.4600465942662324e+00,
       4.4439378615367849e-03, -3.3176117502153640e-03,
       -1.0540748855593554e+01 ]

第二步：估计

执行命令：

./utils/aruco_pose ../utils_calibration/aruco_calibration_grid_board_a4.yml

参考程序：

#include <opencv2/highgui.hpp>
#include "aruco.h"
using namespace std;

int main(int argc,char **argv)
{
 aruco::CameraParameters camera;
 camera.readFromXMLFile(argv[1]);
 aruco::MarkerDetector Detector;
 Detector.setDictionary("ARUCO_MIP_36h12");

 
 cv::namedWindow( "ArUco-Pose-Estimation", cv::WINDOW_AUTOSIZE );
 cv::VideoCapture cap;
 cap.open(0); // open the first camera
 if( !cap.isOpened() ) 
 { // check if we succeeded
  std::cerr << "Couldn't open capture." << std::endl;
  return -1;
 }

 cv::Mat im;
 for (;;)
 {
  cap >> im;
  if ( im.empty() )
   break; // Ran out of film

  auto markers=Detector.detect(im,camera,0.04);//0.102 is the marker size
  for ( auto m:markers )
  {
   aruco::CvDrawingUtils::draw3dAxis(im,m,camera);
   // cout<<m.Rvec<<" "<<m.Tvec<<endl;
             // 实时输出位置坐标 XYZ：
   cout<<m.Tvec<<endl;
  }
  cv::imshow("ArUco-Pose-Estimation",im);
  if ( (char) cv::waitKey(33) >= 0 )
   break;
 }
}