Created by

6330440921 Bhuribhat Ratansasanguanvongs

6330076121 Chanawat Voravijitchaikul

6330585221 Akkharawat Burachokviwat

Medium Link: https://medium.com/@EarthAkkharawat/real-time-vehicle-detection-and-counting-using-the-background-subtractor-comparing-with-haar-ff05f9ab2b92

Download Car Images

!wget https://miro.medium.com/max/640/1*0YyilCiA-lKLNNTZ1QO6rw.webp -O car22.jpeg
!wget https://miro.medium.com/max/640/1*N_FUmoCqT98ZwWVLxdeZZg.webp -O car29.jpeg

Download Car Classifier Model

# cars.xml file from https://github.com/afzal442/Real-Time_Vehicle_Detection-as-Simple/blob/master/cars.xml
!wget https://raw.githubusercontent.com/afzal442/Real-Time_Vehicle_Detection-as-Simple/master/cars.xml -O cars.xml

Detect Car in Images

using cars.xml model

import cv2
import matplotlib.pyplot as plt

# Colors
GREEN = (0, 255, 0)

# Configuration Car Size
min_width  = 30
min_height = 30

# Create Car Classifier
CLF = cv2.CascadeClassifier('cars.xml')

def detect_car_model(image_bgr):
    result = cv2.cvtColor(image_bgr, cv2.COLOR_BGR2RGB)
    gray   = cv2.cvtColor(image_bgr, cv2.COLOR_BGR2GRAY)
    blur   = cv2.GaussianBlur(gray, (3, 3), 5)

    # Pass frame to our car classifier
    cars = CLF.detectMultiScale(
        blur, 
        scaleFactor = 1.1,    # how much the image size is reduced at each image scale
        minNeighbors = 2,     # how many neighbors each candidate rectangle should have to retain it
        minSize = (min_width, min_height)
    )

    # Extract bounding boxes for any car identified
    for x, y, w, h in cars:
        cv2.rectangle(result, (x, y), (x + w, y + h), GREEN, 2)

    plt.imshow(result)
    plt.title(f"Vehicle Count: {len(cars)}")
    plt.show()

Test Function

image = cv2.imread('car29.jpeg')
detect_car_model(image)

image = cv2.imread('car22.jpeg')
detect_car_model(image)

Result

Image 1	Image 2
Car22.jpeg	Car29.jpeg

Real-time Car Detection

usage

>> py main.py

• detect using subtract background (more accuracy)
• detect using cars.xml model (less accuracy)

Resources

• Image dataset
• detect multiscale using opencv
• background subtraction