Skip to content

Interfacing Intel RealSense F200 with OpenCV

RealSense from Intel is an interesting technology that brings 3D cameras into the mainstream.
Although the RealSense SDK provides a good starting point for many applications, some users would prefer to have a bit more control over the images. In this post, I’ll describe how to access the raw streams from an Intel RealSense F200 camera, and how to convert those images into OpenCV cv::Mat objects.

Here is the video with all the explanations

And here are the files used in the video:

01-alignedSingleThreaded.cpp:

#include <pxcsensemanager.h>
#include <opencv2/opencv.hpp>

PXCSenseManager *pxcSenseManager;

PXCImage * CVMat2PXCImage(cv::Mat cvImage)
{
    PXCImage::ImageInfo iinfo;
    memset(&iinfo,0,sizeof(iinfo));
    iinfo.width=cvImage.cols;
    iinfo.height=cvImage.rows;

    PXCImage::PixelFormat format;
    int type = cvImage.type();
    if(type == CV_8UC1)
        format = PXCImage::PIXEL_FORMAT_Y8;
    else if(type == CV_8UC3)
        format = PXCImage::PIXEL_FORMAT_RGB24;
    else if(type == CV_32FC1)
        format = PXCImage::PIXEL_FORMAT_DEPTH_F32;

    iinfo.format = format;

    PXCImage *pxcImage = pxcSenseManager->QuerySession()->CreateImage(&iinfo);

    PXCImage::ImageData data;
    pxcImage->AcquireAccess(PXCImage::ACCESS_WRITE, format, &data);

    data.planes[0] = cvImage.data;

    pxcImage->ReleaseAccess(&data);
    return pxcImage;
}

cv::Mat PXCImage2CVMat(PXCImage *pxcImage, PXCImage::PixelFormat format)
{
    PXCImage::ImageData data;
    pxcImage->AcquireAccess(PXCImage::ACCESS_READ, format, &data);

    int width = pxcImage->QueryInfo().width;
    int height = pxcImage->QueryInfo().height;
    if(!format)
        format = pxcImage->QueryInfo().format;

    int type;
    if(format == PXCImage::PIXEL_FORMAT_Y8)
        type = CV_8UC1;
    else if(format == PXCImage::PIXEL_FORMAT_RGB24)
        type = CV_8UC3;
    else if(format == PXCImage::PIXEL_FORMAT_DEPTH_F32)
        type = CV_32FC1;
    else if(format == PXCImage::PIXEL_FORMAT_DEPTH)
        type = CV_16UC1;

    cv::Mat ocvImage = cv::Mat(cv::Size(width, height), type, data.planes[0]);

    pxcImage->ReleaseAccess(&data);
    return ocvImage;
}

int main(int argc, char* argv[])
{
    //Define some parameters for the camera
    cv::Size frameSize = cv::Size(640, 480);
    float frameRate = 60;

    //Create the OpenCV windows and images
    cv::namedWindow("IR", cv::WINDOW_NORMAL);
    cv::namedWindow("Color", cv::WINDOW_NORMAL);
    cv::namedWindow("Depth", cv::WINDOW_NORMAL);
    cv::Mat frameIR = cv::Mat::zeros(frameSize, CV_8UC1);
    cv::Mat frameColor = cv::Mat::zeros(frameSize, CV_8UC3);
    cv::Mat frameDepth = cv::Mat::zeros(frameSize, CV_8UC1);

    //Initialize the RealSense Manager
    pxcSenseManager = PXCSenseManager::CreateInstance();

    //Enable the streams to be used
    pxcSenseManager->EnableStream(PXCCapture::STREAM_TYPE_IR, frameSize.width, frameSize.height, frameRate);
    pxcSenseManager->EnableStream(PXCCapture::STREAM_TYPE_COLOR, frameSize.width, frameSize.height, frameRate);
    pxcSenseManager->EnableStream(PXCCapture::STREAM_TYPE_DEPTH, frameSize.width, frameSize.height, frameRate);

    //Initialize the pipeline
    pxcSenseManager->Init();

    bool keepRunning = true;
    while(keepRunning)
    {
        //Acquire all the frames from the camera
        pxcSenseManager->AcquireFrame();
        PXCCapture::Sample *sample = pxcSenseManager->QuerySample();

        //Convert each frame into an OpenCV image
        frameIR = PXCImage2CVMat(sample->ir, PXCImage::PIXEL_FORMAT_Y8);
        frameColor = PXCImage2CVMat(sample->color, PXCImage::PIXEL_FORMAT_RGB24);
        cv::Mat frameDepth_u16 = PXCImage2CVMat(sample->depth, PXCImage::PIXEL_FORMAT_DEPTH);
        frameDepth_u16.convertTo(frameDepth, CV_8UC1);

        cv::Mat frameDisplay;
        cv::equalizeHist(frameDepth, frameDisplay);

        //Display the images
        cv::imshow("IR", frameIR);
        cv::imshow("Color", frameColor);
        cv::imshow("Depth", frameDisplay);

        //Check for user input
        int key = cv::waitKey(1);
        if(key == 27)
            keepRunning = false;

        //Release the memory from the frames
        pxcSenseManager->ReleaseFrame();
    }

    //Release the memory from the RealSense manager
    pxcSenseManager->Release();

    return 0;
}

02-unalignedSingleThreaded.cpp:
#include <pxcsensemanager.h>
#include <opencv2/opencv.hpp>

cv::Mat PXCImage2CVMat(PXCImage *pxcImage, PXCImage::PixelFormat format)
{
    PXCImage::ImageData data;
    pxcImage->AcquireAccess(PXCImage::ACCESS_READ, format, &data);

    int width = pxcImage->QueryInfo().width;
    int height = pxcImage->QueryInfo().height;

    if(!format)
        format = pxcImage->QueryInfo().format;

    int type;
    if(format == PXCImage::PIXEL_FORMAT_Y8)
        type = CV_8UC1;
    else if(format == PXCImage::PIXEL_FORMAT_RGB24)
        type = CV_8UC3;
    else if(format == PXCImage::PIXEL_FORMAT_DEPTH_F32)
        type = CV_32FC1;

    cv::Mat ocvImage = cv::Mat(cv::Size(width, height), type, data.planes[0]);

    pxcImage->ReleaseAccess(&data);
    return ocvImage;
}

int main(int argc, char* argv[])
{
    //Define some parameters for the camera
    cv::Size frameSize = cv::Size(640, 480);
    float frameRate = 60;

    //Create the OpenCV windows and images
    cv::namedWindow("IR", cv::WINDOW_NORMAL);
    cv::namedWindow("Color", cv::WINDOW_NORMAL);
    cv::namedWindow("Depth", cv::WINDOW_NORMAL);
    cv::Mat frameIR = cv::Mat::zeros(frameSize, CV_8UC1);
    cv::Mat frameColor = cv::Mat::zeros(frameSize, CV_8UC3);
    cv::Mat frameDepth = cv::Mat::zeros(frameSize, CV_8UC1);

    //Initialize the RealSense Manager
    PXCSenseManager *pxcSenseManager = PXCSenseManager::CreateInstance();

    //Enable the streams to be used
    pxcSenseManager->EnableStream(PXCCapture::STREAM_TYPE_IR, frameSize.width, frameSize.height, frameRate);
    pxcSenseManager->EnableStream(PXCCapture::STREAM_TYPE_COLOR, frameSize.width, frameSize.height, frameRate);
    pxcSenseManager->EnableStream(PXCCapture::STREAM_TYPE_DEPTH, frameSize.width, frameSize.height, frameRate);

    //Initialize the pipeline
    pxcSenseManager->Init();

    bool keepRunning = true;
    while(keepRunning)
    {
        //Acquire any frame from the camera
        pxcSenseManager->AcquireFrame(false);
        PXCCapture::Sample *sample = pxcSenseManager->QuerySample();

        //Convert each frame into an OpenCV image
        //You need to make sure that the image is there first
        if(sample->ir)
            frameIR = PXCImage2CVMat(sample->ir, PXCImage::PIXEL_FORMAT_Y8);
        if(sample->color)
            frameColor = PXCImage2CVMat(sample->color, PXCImage::PIXEL_FORMAT_RGB24);
        if(sample->depth)
            PXCImage2CVMat(sample->depth, PXCImage::PIXEL_FORMAT_DEPTH_F32).convertTo(frameDepth, CV_8UC1);

        //Display the images
        cv::imshow("IR", frameIR);
        cv::imshow("Color", frameColor);
        cv::imshow("Depth", frameDepth);

        //Check for user input
        int key = cv::waitKey(1);
        if(key == 27)
            keepRunning = false;

        //Release the memory from the frames
        pxcSenseManager->ReleaseFrame();
    }

    //Release the memory from the RealSense manager
    pxcSenseManager->Release();

    return 0;
}

03-unalignedMultiThreaded.cpp:
#include <pxcsensemanager.h>
#include <iostream>
#include <opencv2/opencv.hpp>

cv::Mat frameIR;
cv::Mat frameColor;
cv::Mat frameDepth;
cv::Mutex framesMutex;

cv::Mat PXCImage2CVMat(PXCImage *pxcImage, PXCImage::PixelFormat format)
{
    PXCImage::ImageData data;
    pxcImage->AcquireAccess(PXCImage::ACCESS_READ, format, &data);

    int width = pxcImage->QueryInfo().width;
    int height = pxcImage->QueryInfo().height;

    if(!format)
        format = pxcImage->QueryInfo().format;

    int type;
    if(format == PXCImage::PIXEL_FORMAT_Y8)
        type = CV_8UC1;
    else if(format == PXCImage::PIXEL_FORMAT_RGB24)
        type = CV_8UC3;
    else if(format == PXCImage::PIXEL_FORMAT_DEPTH_F32)
        type = CV_32FC1;

    cv::Mat ocvImage = cv::Mat(cv::Size(width, height), type, data.planes[0]);

    pxcImage->ReleaseAccess(&data);
    return ocvImage;
}

class FramesHandler:public PXCSenseManager::Handler
{
public:
    virtual pxcStatus PXCAPI OnNewSample(pxcUID, PXCCapture::Sample *sample)
    {
            framesMutex.lock();
                if(sample->ir)
                    frameIR = PXCImage2CVMat(sample->ir, PXCImage::PIXEL_FORMAT_Y8);
                if(sample->color)
                    frameColor = PXCImage2CVMat(sample->color, PXCImage::PIXEL_FORMAT_RGB24);
                if(sample->depth)
                    PXCImage2CVMat(sample->depth, PXCImage::PIXEL_FORMAT_DEPTH_F32).convertTo(frameDepth, CV_8UC1);
            framesMutex.unlock();
    return PXC_STATUS_NO_ERROR;
    }
};

int main(int argc, char* argv[])
{
    cv::Size frameSize = cv::Size(640, 480);
    float frameRate = 60;

    cv::namedWindow("IR", cv::WINDOW_NORMAL);
    cv::namedWindow("Color", cv::WINDOW_NORMAL);
    cv::namedWindow("Depth", cv::WINDOW_NORMAL);
    frameIR = cv::Mat::zeros(frameSize, CV_8UC1);
    frameColor = cv::Mat::zeros(frameSize, CV_8UC3);
    frameDepth = cv::Mat::zeros(frameSize, CV_8UC1);

    PXCSenseManager *pxcSenseManager = PXCSenseManager::CreateInstance();

    //Enable the streams to be used
    pxcSenseManager->EnableStream(PXCCapture::STREAM_TYPE_IR, frameSize.width, frameSize.height, frameRate);
    pxcSenseManager->EnableStream(PXCCapture::STREAM_TYPE_COLOR, frameSize.width, frameSize.height, frameRate);
    pxcSenseManager->EnableStream(PXCCapture::STREAM_TYPE_DEPTH, frameSize.width, frameSize.height, frameRate);

    FramesHandler handler;
    pxcSenseManager->Init(&handler);
    pxcSenseManager->StreamFrames(false);

    //Local images for display
    cv::Mat displayIR = frameIR.clone();
    cv::Mat displayColor = frameColor.clone();
    cv::Mat displayDepth = frameDepth.clone();

    bool keepRunning = true;
    while(keepRunning)
    {
        framesMutex.lock();
            displayIR = frameIR.clone();
            displayColor = frameColor.clone();
            displayDepth = frameDepth.clone();
        framesMutex.unlock();

        cv::imshow("IR", displayIR);
        cv::imshow("Color", displayColor);
        cv::imshow("Depth", displayDepth);

        int key = cv::waitKey(1);
        if(key == 27)
            keepRunning = false;
    }
    //Stop the frame acqusition thread
    pxcSenseManager->Close();

    pxcSenseManager->Release();

    return 0;
}

04-alignedMultiThreaded.cpp:
#include <pxcsensemanager.h>
#include <iostream>
#include <opencv2/opencv.hpp>

cv::Mat frameIR;
cv::Mat frameColor;
cv::Mat frameDepth;
cv::Mutex framesMutex;

cv::Mat PXCImage2CVMat(PXCImage *pxcImage, PXCImage::PixelFormat format)
{
    PXCImage::ImageData data;
    pxcImage->AcquireAccess(PXCImage::ACCESS_READ, format, &data);

    int width = pxcImage->QueryInfo().width;
    int height = pxcImage->QueryInfo().height;

    if(!format)
        format = pxcImage->QueryInfo().format;

    int type;
    if(format == PXCImage::PIXEL_FORMAT_Y8)
        type = CV_8UC1;
    else if(format == PXCImage::PIXEL_FORMAT_RGB24)
        type = CV_8UC3;
    else if(format == PXCImage::PIXEL_FORMAT_DEPTH_F32)
        type = CV_32FC1;

    cv::Mat ocvImage = cv::Mat(cv::Size(width, height), type, data.planes[0]);

    pxcImage->ReleaseAccess(&data);
    return ocvImage;
}

class FramesHandler:public PXCSenseManager::Handler
{
public:
    virtual pxcStatus PXCAPI OnNewSample(pxcUID, PXCCapture::Sample *sample)
    {
            framesMutex.lock();
                frameIR = PXCImage2CVMat(sample->ir, PXCImage::PIXEL_FORMAT_Y8);
                frameColor = PXCImage2CVMat(sample->color, PXCImage::PIXEL_FORMAT_RGB24);
                PXCImage2CVMat(sample->depth, PXCImage::PIXEL_FORMAT_DEPTH_F32).convertTo(frameDepth, CV_8UC1);
            framesMutex.unlock();
    return PXC_STATUS_NO_ERROR;
    }
};

int main(int argc, char* argv[])
{
    cv::Size frameSize = cv::Size(640, 480);
    float frameRate = 60;

    cv::namedWindow("IR", cv::WINDOW_NORMAL);
    cv::namedWindow("Color", cv::WINDOW_NORMAL);
    cv::namedWindow("Depth", cv::WINDOW_NORMAL);
    frameIR = cv::Mat::zeros(frameSize, CV_8UC1);
    frameColor = cv::Mat::zeros(frameSize, CV_8UC3);
    frameDepth = cv::Mat::zeros(frameSize, CV_8UC1);

    PXCSenseManager *pxcSenseManager = PXCSenseManager::CreateInstance();

    //Enable the streams to be used
    PXCVideoModule::DataDesc ddesc={};
    ddesc.deviceInfo.streams = PXCCapture::STREAM_TYPE_IR | PXCCapture::STREAM_TYPE_COLOR | PXCCapture::STREAM_TYPE_DEPTH;

    pxcSenseManager->EnableStreams(&ddesc);

    FramesHandler handler;
    pxcSenseManager->Init(&handler);
    pxcSenseManager->StreamFrames(false);

    //Local images for display
    cv::Mat displayIR = frameIR.clone();
    cv::Mat displayColor = frameColor.clone();
    cv::Mat displayDepth = frameDepth.clone();

    bool keepRunning = true;
    while(keepRunning)
    {
        framesMutex.lock();
            displayIR = frameIR.clone();
            displayColor = frameColor.clone();
            displayDepth = frameDepth.clone();
        framesMutex.unlock();

        cv::imshow("IR", displayIR);
        cv::imshow("Color", displayColor);
        cv::imshow("Depth", displayDepth);

        int key = cv::waitKey(1);
        if(key == 27)
            keepRunning = false;
    }
    //Stop the frame acqusition thread
    pxcSenseManager->Close();

    pxcSenseManager->Release();

    return 0;
}

Posted in Computer Vision, IoT, OpenCV, Photography, Programming.

rev="post-404" 12 comments

By samontab – April 27, 2016

12 Responses

Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.

  1. Jonathan T. says

    In your conversion functions, you need to add handling for the “pitches” of the images, which are equivalent to the cv::Mat’s “step”. So when converting to PXC, after data.planes[0] = cvImage.data, you should set data.pitches[0] = cvImage.step, and when converting to Mat, you should use the Mat constructor that passes the step like this: cv::Mat ocvImage = cv::Mat(cv::Size(width, height), type, data.planes[0], data.pitches[0]);

    Good luck!

    June 9, 2016, 9:22 pm
  2. samontab says

    You’re right Jonathan, thanks!

    June 9, 2016, 9:28 pm
  3. John M. says

    do the *.cpp files above work alone without any header? i am trying to use this on r200 and it says missing stdafx.h

    if i include it, the program just terminates directly after start.

    The thread 0x1770 has exited with code -1073741701 (0xc000007b).
    The program ‘[2744] RealSenseUnalignedSingleThread.exe’ has exited with code -1073741701 (0xc000007b).

    June 27, 2016, 9:43 am
  4. samontab says

    You shouldn’t need any precompiled header.

    July 7, 2016, 1:13 am
  5. dude says

    Hi which openCV version were you using?

    December 30, 2016, 8:17 pm
  6. samontab says

    It should work with the latest versions, either the 2.x or 3.x branch.

    January 7, 2017, 10:24 am
  7. sheom says

    Hello, Sebastian
    i have some questions, but its not focus on this poster. Sorry.
    I get two 3D coordinates by using realSense 200 Camera.
    And want to measure between two coordinates.
    i can measure distance from camera to object. but i cannot find way to measure b/w two 3D coordinate.
    Help me please T.T

    January 24, 2017, 4:20 am
  8. samontab says

    Hey sheom,

    Just get the euclidean distance, i.e. sqrt((x2 – x1)^2 + (y2 – y1)^2 + (z2 – z1)^2)

    January 30, 2017, 5:17 am
  9. Ron says

    Code to interact with UVC camera extension unit controls?

    I have an `Intel RealSense 410 (DS5)` 3D camera, and I would like to manipulate its LASER *extended control*.
    D’you knows of code snippets* demonstrating how to interact with UVC extension unit controls – using OpenCV?
    —
    * Preferably in Python.

    February 23, 2017, 8:17 am
  10. samontab says

    Hi Ron,

    OpenCV is independent to the camera acquisition or control. You should be able to control the camera using librealsense.

    March 12, 2017, 10:45 pm
  11. Lucie says

    Hello samontab,

    I am working with the R200 camera and I followed your video.
    Unfortunately, I got a black and white image instead of the grey you obtain for the depth stream when I am converting it in cv::Mat.
    Have you ever had the same problem?
    Thanks,
    Lucie

    April 27, 2017, 11:02 am
  12. samontab says

    Lucie, you should be able to get a similar depth image from the R200.

    Make sure you are requesting the correct format for R200,and that you are using the corresponding data type in OpenCV.
    Also, have a look at librealsense for an official cross platform library to access the camera data

    April 28, 2017, 2:22 am



Some HTML is OK

or, reply to this post via trackback.

« »