Vehicle detection in Python using OpenCV

In this tutorial, we will learn how to detect multiple vehicles in the same frame. We will use the OpenCV Python library to perform this task. Install Python OpenCV library on your Python environment if you have not installed it yet.

Secondly, we need the haar cascaded files which contains the features of the vehicles which we want to detect. Here we are using cars.xml and bicycle.xml.

This program will detect the cars and the bicycles from the given footage and draws rectangle over it.

Now we will proceed towards our coding section.

Python code Vehicle detection using OpenCV

Let’s start our code:

import cv2

Import Python OpenCV library.

cap = cv2.VideoCapture('traffic.mp4')  #Path to footage
car_cascade = cv2.CascadeClassifier('cars.xml')  #Path to cars.xml
bicycle_cascade= cv2.CascadeClassifier('bicycle.xml') #Path to bicycle.xml

Use the VideoCapture method to process the video or footage. Give the path of vehicle models to the CascadeClassifier function.

while True:
    ret, img = cap.read()
    gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
    cars=car_cascade.detectMultiScale(gray,1.8,2)
    bicycle = bicycle_cascade.detectMultiScale(gray, 1.8, 2)

cap.read() function reads every frame and gives an img array. cvtColor changes image into grayscale. We do not need a colored frame to detect objects moreover it decreases the processing speed. cars and bicycle stores (x,y,w,h) for detected objects. (x,y) are location and w stands for width and h for height.

    #Drawing rectangles on detected cars
    for (x,y,w,h) in cars:
        cv2.rectangle(img,(x,y),(x+w,y+h),(225,0,0),2)

    # Drawing rectangles on detected bicycles
    for (x,y,w,h) in bicycle:
        cv2.rectangle(img,(x,y),(x+w,y+h),(225,0,0),2)

For every (x,y,w,h) in cars or bicycle cv2.rectangle line draws a rectangle. The first argument of the rectangle is image on which we will draw the rectangle. Another two are associated with coordinates. Then color and after it the width of a line.

    cv2.imshow('img',img) #Shows the frame

Enter imshow line to display the image with the label ‘img’.

    if cv2.waitKey(20) & 0xFF == ord('q'):
        break

If we press ‘q’ on the keyboard then the program will terminate.

cap.release()
cv2.destroyAllWindows()

Then come out of while loop and write the lines described in the upper code block. These lines will release the camera device if any and de-allocate memory which was used earlier.

Now, have a look at the below box which contains the full code of this given problem.

import cv2

cap = cv2.VideoCapture('traffic.mp4')  #Path to footage
car_cascade = cv2.CascadeClassifier('cars.xml')  #Path to cars.xml
bicycle_cascade= cv2.CascadeClassifier('bicycle.xml') #Path to bicycle.xml

while True:
    ret, img = cap.read()
    gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
    cars=car_cascade.detectMultiScale(gray,1.8,2)
    bicycle = bicycle_cascade.detectMultiScale(gray, 1.8, 2)

    #Drawing rectangles on detected cars
    for (x,y,w,h) in cars:
        cv2.rectangle(img,(x,y),(x+w,y+h),(225,0,0),2)

    # Drawing rectangles on detected bicycles
    for (x,y,w,h) in bicycle:
        cv2.rectangle(img,(x,y),(x+w,y+h),(225,0,0),2)


    cv2.imshow('img',img) #Shows the frame



    if cv2.waitKey(20) & 0xFF == ord('q'):
        break

cap.release()
cv2.destroyAllWindows()

If everything goes fine then it will show the blue rectangles around detected cars and bicycles for given footage. So in this way, we have successfully detected vehicles using Python OpenCV.