simple way to get Z "depth" in OpenCV - c++

Question

I need to detect a blue object and a red object from two different cameras, the required task for now is to locate each object position in 3D space, this means for each object we have to have its x,y,z coordinates, I've seen this video and here witch does exactly what i'm trying to do but there was no sample code in case of the first video, my code looks like this for now it gets me x,y of red/blue object but no depth:
#include <iostream>
#include "opencv2/highgui/highgui.hpp"
#include "opencv2/imgproc/imgproc.hpp"
#include <opencv/highgui.h>
using namespace cv;
using namespace std;
int main(int argc, char** argv)
//int func(int argc, char** argv)
{
VideoCapture cap(0); //capture the video from webcam
VideoCapture cap1(1); //capture the video from extrenal camers
if (!cap.isOpened()) // if not success, exit program
{
cout << "Cannot open the web cam" << endl;
return -1;
}
if (!cap1.isOpened()) // if not success, exit program
{
cout << "Cannot open the External camera" << endl;
return -1;
}
namedWindow("Control", CV_WINDOW_AUTOSIZE); //create a window called "Control"
int iLowH = 170;
int iHighH = 179;
int iLowS = 150;
int iHighS = 255;
int iLowV = 60;
int iHighV = 255;
//Create trackbars in "Control" window to control the range of red detection
createTrackbar("LowH", "Control", &iLowH, 179); //Hue (0 - 179)
createTrackbar("HighH", "Control", &iHighH, 179);
createTrackbar("LowS", "Control", &iLowS, 255); //Saturation (0 - 255)
createTrackbar("HighS", "Control", &iHighS, 255);
createTrackbar("LowV", "Control", &iLowV, 255);//Value (0 - 255)
createTrackbar("HighV", "Control", &iHighV, 255);
int iLastX = -1; //last known co-ordinates of red object
int iLastY = -1;
int iLastX1 = -1;
int iLastY1 = -1;
//Capture a temporary image from both cameras to obtain size
Mat imgTmp;
cap.read(imgTmp);
cap1.read(imgTmp);
//Create a black image with the size as the camera output
Mat imgLines = Mat::zeros(imgTmp.size(), CV_8UC3);;
Mat imgLines1 = Mat::zeros(imgTmp.size(), CV_8UC3);;
//loop of continuously capturing frames from video
while (true)
{
Mat imgOriginal;
Mat imgOriginal1;
bool bSuccess = cap.read(imgOriginal); // read a new frame from video webcam
bool bSuccess1 = cap1.read(imgOriginal1); // read a new frame from video external cam
if (!bSuccess || !bSuccess1) //if not success, break loop
{
cout << "Cannot read a frame from video stream" << endl;
break;
}
//WebCam code for image and tracking/detecting
Mat imgHSV;
cvtColor(imgOriginal, imgHSV, COLOR_BGR2HSV); //Convert the captured frame from BGR to HSV to control range of color to obtain and be able to detect it
Mat imgThresholded;
inRange(imgHSV, Scalar(iLowH, iLowS, iLowV), Scalar(iHighH, iHighS, iHighV), imgThresholded); //Threshold the image at the colors within specified range
//morphological opening (removes noise and similar colored objects appearing in thresholded image)
erode(imgThresholded, imgThresholded, getStructuringElement(MORPH_ELLIPSE, Size(5, 5)));
dilate(imgThresholded, imgThresholded, getStructuringElement(MORPH_ELLIPSE, Size(5, 5)));
//morphological closing (removes noise appearing inside our object in the thresholded image)
dilate(imgThresholded, imgThresholded, getStructuringElement(MORPH_ELLIPSE, Size(5, 5)));
erode(imgThresholded, imgThresholded, getStructuringElement(MORPH_ELLIPSE, Size(5, 5)));
//Calculate the moments of the thresholded image to calculate the object position
Moments oMoments = moments(imgThresholded);
double dM01 = oMoments.m01;
double dM10 = oMoments.m10;
double dArea = oMoments.m00;
// if the area <= 10000, I consider that the there are no object in the image and it's because of the noise, the area is not zero
if (dArea > 10000)
{
//calculate the position of the ball
int posX = dM10 / dArea;
int posY = dM01 / dArea;
if (iLastX >= 0 && iLastY >= 0 && posX >= 0 && posY >= 0)
{
//Draw a red line from the previous point to the current point
line(imgLines, Point(posX, posY), Point(iLastX, iLastY), Scalar(0, 0, 255), 2);
}
iLastX = posX; //current point becomes last known point and loop continues
iLastY = posY;
}
imshow("Thresholded Image", imgThresholded); //show the thresholded image
imgOriginal = imgOriginal + imgLines;
imshow("Original", imgOriginal); //show the original image with the tracking lines if exist
//External Cam code track/detect
Mat imgHSV1;
cvtColor(imgOriginal1, imgHSV1, COLOR_BGR2HSV); //Convert the captured frame from BGR to HSV
Mat imgThresholded1;
inRange(imgHSV1, Scalar(iLowH, iLowS, iLowV), Scalar(iHighH, iHighS, iHighV), imgThresholded1); //Threshold the image
//morphological opening (removes noise and similar colored objects appearing in thresholded image)
erode(imgThresholded1, imgThresholded1, getStructuringElement(MORPH_ELLIPSE, Size(5, 5)));
dilate(imgThresholded1, imgThresholded1, getStructuringElement(MORPH_ELLIPSE, Size(5, 5)));
//morphological closing (removes noise appearing inside our object in the thresholded image)
dilate(imgThresholded1, imgThresholded1, getStructuringElement(MORPH_ELLIPSE, Size(5, 5)));
erode(imgThresholded1, imgThresholded1, getStructuringElement(MORPH_ELLIPSE, Size(5, 5)));
//Calculate the moments of the thresholded image to calculate the object position
Moments oMoments1 = moments(imgThresholded1);
double dM011 = oMoments1.m01;
double dM101 = oMoments1.m10;
double dArea1 = oMoments1.m00;
// if the area <= 10000, I consider that the there are no object in the image and it's because of the noise, the area is not zero
if (dArea1 > 10000)
{
//calculate the position of the ball
int posX1 = dM101 / dArea1;
int posY1 = dM011 / dArea1;
if (iLastX1 >= 0 && iLastY1 >= 0 && posX1 >= 0 && posY1 >= 0)
{
//Draw a red line from the previous point to the current point
line(imgLines1, Point(posX1, posY1), Point(iLastX1, iLastY1), Scalar(0, 0, 255), 2);
}
iLastX1 = posX1;
iLastY1 = posY1;
}
imshow("Thresholded Image 2", imgThresholded1); //show the thresholded image
imgOriginal1 = imgOriginal1 + imgLines1;
imshow("Original 2", imgOriginal1); //show the original image
if (waitKey(30) == 27) //wait for 'esc' key press for 30ms. If 'esc' key is pressed, break loop
{
cout << "esc key is pressed by user" << endl;
break;
}
}
return 0;
}

Answer 1

The answer to your question will be Stereo Vision. You need to do the Stereo Calibration of the two cameras in order to obtain the transformation matrix that allows to produce the depth map of the scene from the 2 views. OpenCV provides some functions to do that.
Here is a tutorial to begin with.