This project detects green-colored objects in a live video feed, creates bounding boxes around them, and filters out noise using NumPy for computational tasks and OpenCV for video processing.
- Detects green objects using a custom mask creation function with NumPy.
- Filters small objects based on area threshold to remove noise.
- Draws bounding boxes around each detected object using NumPy, without relying on
cv2.findContours. - Real-time video feed processing with a webcam.
Ensure you have the following installed:
- Python 3.8 or later
- OpenCV
- NumPy
-
Clone the repository:
git clone https://github.com/VishvendraTomar/green-object-detection.git cd green-object-detection -
Install the required Python packages:
pip install opencv-python numpy
-
Run the program:
python main.py
-
Allow the webcam to start. The program will detect green objects and draw bounding boxes around them.
-
Press
qto quit the application.
.
├── main.py # Main script to run the detection
├── requirements.txt # Dependencies list
└── README.md # Project documentation
create_mask: Generates a binary mask to detect green pixels using NumPy.find_bounding_boxes: Identifies separate green objects and calculates bounding boxes without usingcv2.findContours.- Real-time Processing: Captures frames from a webcam, processes them in real-time, and displays the results.
The program processes frames from a live video feed, identifying multiple green objects and bounding them with boxes. It uses area thresholds to ignore small objects (e.g., noise).
- HSV Range: Modify
lower_hue,upper_hue,lower_saturation, andupper_saturationto adjust the detection for other colors or different lighting conditions. - Minimum Area: Adjust the
min_areaparameter infind_bounding_boxesto detect smaller or larger objects.