Live camera for non-ROS build

[originholic]

This is probably a fundamental issues, I am wondering that is it possible to run a live camera with SVO that built without ROS? and how can I modify the code to do so? Many thanks

[inakimalerba]

Hi @originholic , are you still working on ths?

[GhisHD]

Hi, Current code should works without (wiki give a method), I currently run it on MINGW on windows without ROS and using camera input from opencv method.

Thks Ghis

[YoshuaNava]

@GhisHD, excuse me. Which opencv method are you using to feed rpg_svo with camera images?

Thank you very much. Yoshua

[xxFirefly]

@YoshuaNava i think he meant just rewriting function runFromFolder in test_pipeline.cpp, by changing source of images. For getting images from camera you should create opencv camera: cv::VideoCapture cap(0); than in process cycle read img from camera and convert it to grayscale:

bool bSuccess = cap.read(img);
cv::cvtColor(img, img, CV_RGB2GRAY);

and add it to svo vo_->addImage(img, 0.01*img_id);

For example:

void BenchmarkNode::runFromCamera()
{
  cv::VideoCapture cap(0);

  if (!cap.isOpened())  // if not success, exit program
  {
    std::cout << "Cannot open the video cam" << endl;
    return;
  }

  double dWidth = cap.get(CV_CAP_PROP_FRAME_WIDTH); //get the width of frames of the video
  double dHeight = cap.get(CV_CAP_PROP_FRAME_HEIGHT); //get the height of frames of the video

  std::cout << "Frame size : " << dWidth << " x " << dHeight << endl;

  for (int img_id = 2;;++img_id)
  {
    cv::Mat img;

    bool bSuccess = cap.read(img);
    cv::cvtColor(img, img, CV_RGB2GRAY);

    if (!bSuccess)
    {
        std::cout << "Cannot read a frame from video stream" << endl;

    }

    cv::imshow("Video stream", img);

    if (cv::waitKey(30) == 27)
    {
        cout << "esk key is pressed" << endl;
        break;
    }

    assert(!img.empty());

    // process frame
    vo_->addImage(img, 0.01*img_id);

    SE3 T_world_from_vision_;
    SE3 T_world_from_cam;
    Vector3d p; 

    T_world_from_vision_ = SE3(Matrix3d::Identity(), Vector3d::Zero());
    T_world_from_cam = SE3(T_world_from_vision_*vo_->lastFrame()->T_f_w_.inverse());
    p = Vector3d(vo_->lastFrame()->T_f_w_.translation());

    // display tracking quality
    if(vo_->lastFrame() != NULL)
    {
//          std::cout << "Frame-Id: " << vo_->lastFrame()->id_ << " \t"
            std::cout << "#Features: " << vo_->lastNumObservations() << " \t";
//                    << "Proc. Time: " << vo_->lastProcessingTime()*1000 << "ms \t"
        std::cout << "X: " << p[0] << " \t"
                  << "Y: " << p[1] << " \t"
                  << "Z: " << p[2] << "\n";

        // access the pose of the camera via vo_->lastFrame()->T_f_w_.
    }

  }
}

[rahul95ram]

Hi guys, were you able to get SVO working on a live camera without ROS? Im having trouble getting it working. Any help would be appreciated!

[rahul95ram]

Hi guys, how do I get the live camera view as shown in the video. I want to use svo on a raspberry pi3 and track the path traversed by my mobile robot. After compiling and building the package, I ran the svo ( without ros) test_pipeline. The programme ran, but I was not able to see the output tracking.

Any help would be appreciated thanks!

[chrissunny94]

@lorhm , Were you successful in running it on the PI?