Static image vehicle recognition using opencv c++

Question

I am working on a project of using a camera to take a picture for detecting the available parking space in the carpark.

I am going to use image subtraction for comparing the empty and occupied situation in the carpark.

These are the two picture for comparison

This is the resulting image:

Can anyone tell me how to

recognize and put a rectangle outside the vehicle
output the number of the car for my further usage?

Here is the code:

    // Test.cpp

#include<opencv2/core/core.hpp>
#include<opencv2/highgui/highgui.hpp>
#include<opencv2/imgproc/imgproc.hpp>

#include<iostream>

using namespace std;
using namespace cv;

Mat bg_frame;
Mat cam_frame;
Mat diff_frame;
char charCheckForEscKey = 0;

int main() {

    while (charCheckForEscKey != 27 /*&& capWebcam.isOpened()*/) {

    bg_frame = imread("Picture1.jpg",0);
    cam_frame = imread("Picture4.jpg",0);

    resize(cam_frame, cam_frame, bg_frame.size());

    imshow("test1", bg_frame);
    imshow("test2", cam_frame);

    absdiff(bg_frame, cam_frame, diff_frame);

    threshold(diff_frame, diff_frame, 80, 255, THRESH_BINARY);


    //erode(diff_frame, diff_frame, getStructuringElement(MORPH_RECT, Size(1, 1)));

    dilate(diff_frame, diff_frame, Mat(), Point(-1, -1));

    //Canny(diff_frame, diff_frame, 30, 70);

    imshow("test3", diff_frame);

    charCheckForEscKey = cv::waitKey(1);            // delay (in ms) and get key press, if any
    }

    return(0);

}

Hennio · Accepted Answer · 2017-03-10 09:53:54Z

1

Your problem is very complex to be solved in a single post, but I will give you some pointers on how to solve it.

The idea is correct, you should compare the differences between images. I would not recommend comparing directly pixel by pixel, but its a fast solution for a prototype. For more robust methods checkout background subtraction c++ and more extense but in python

As for counting the number of vehicles, given the complexity of the scene (overlapping cars, non uniform parking) then I would recommend manually specifying a region of interest (ROI) for each car. Applying the method you already have for each ROI, you can detect if that parking space is occupied or not.

On how to draw a rectangle in opencv.

void rectangle(Mat& img, Point pt1, Point pt2, const Scalar& color, int thickness=1, int lineType=8, int shift=0);

answered Mar 10, 2017 at 9:53

Hennio

4432 silver badges18 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Edward Wong Over a year ago

Thank you so much, can you talk about more about ROI?

Hennio Over a year ago

its simply a rectangular section of the image, a series of pixels that define a region. In your case its a region per parking space. This way you can split the global problem to a simple case of "is this image a car or an empty park space?"

Collectives™ on Stack Overflow

Static image vehicle recognition using opencv c++

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related