Vehicle detection in a Video frame using Python - OpenCV

Detecting vehicles in a video frame is a common application in computer vision, especially for surveillance, traffic monitoring, and autonomous vehicles. Here, we'll use OpenCV's pre-trained Haar cascades to detect vehicles in a video frame. Although this method is not as robust as deep learning techniques, it's simpler and faster.

Prerequisites:

Install necessary libraries:

pip install opencv-python

Obtain a trained cascade XML file for vehicle (car) detection. OpenCV provides a few Haar cascades (like face and eyes), but you might need to find or train one for vehicles. There are a few cascades available online (e.g., car detection Haar cascades) which you can utilize.

Step-by-Step Tutorial:

Import necessary libraries:

import cv2

Load the pre-trained cascade:

car_cascade = cv2.CascadeClassifier('path_to_vehicle_cascade.xml')

Process the video and detect cars:

# Open video
video = cv2.VideoCapture('path_to_video.mp4')

while True:
    ret, frame = video.read()
    
    if not ret:
        break
    
    # Convert frame to grayscale
    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    
    # Detect cars
    cars = car_cascade.detectMultiScale(gray, 1.1, 3)
    
    # Draw bounding box around cars
    for (x, y, w, h) in cars:
        cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 0), 2)
        
    # Display frame
    cv2.imshow('Car Detection', frame)
    
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

video.release()
cv2.destroyAllWindows()

Limitations and Considerations:

Robustness: Haar cascades are not the most robust technique for car detection, especially in complex scenarios. It might have trouble detecting cars from different angles or in various lighting conditions.
False Positives/Negatives: Depending on the cascade quality and the detection parameters, you might experience false positives (detecting non-car objects as cars) or false negatives (missing some cars).
Deep Learning: For robust vehicle detection in complex scenarios, consider using deep learning techniques. Popular architectures like Single Shot MultiBox Detector (SSD), You Only Look Once (YOLO), and Faster R-CNN have been used effectively for vehicle detection tasks.
Training: If you're unable to find a satisfactory pre-trained cascade for car detection, you can consider training one using positive and negative samples. OpenCV provides tools for training Haar cascades, though this can be a time-consuming process.

Remember, the approach presented here is basic and might not work perfectly in all scenarios. However, it serves as an introduction to the concept of object detection using OpenCV.

Vehicle detection in video using OpenCV Python:

Description: Detecting vehicles in a video stream by applying object detection techniques.

Code:

import cv2

# Load pre-trained vehicle detection model
cascade_path = 'haarcascade_car.xml'
car_cascade = cv2.CascadeClassifier(cascade_path)

# Read video stream
cap = cv2.VideoCapture('traffic_video.mp4')

while True:
    ret, frame = cap.read()

    # Convert to grayscale for Haarcascade
    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

    # Detect vehicles
    cars = car_cascade.detectMultiScale(gray, scaleFactor=1.1, minNeighbors=5)

    # Draw bounding boxes around detected vehicles
    for (x, y, w, h) in cars:
        cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 255, 0), 2)

    # Display the resulting frame
    cv2.imshow('Vehicle Detection', frame)

    if cv2.waitKey(25) & 0xFF == ord('q'):
        break

cap.release()
cv2.destroyAllWindows()