By Konstantinos Patronas —

Object detection using Python and the YOLO library

In this very short article i will show you a very simple example on how to use the YOLO library with pre-trained models, i think its the…

Object detection using Python and the YOLO library
Photo by Markus Winkler on Unsplash

In this very short article i will show you a very simple example on how to use the YOLO library with pre-trained models, i think its the simplest example someone can make with YOLO, but most of the times simple things is what you need to start exploring a new topic!

What is YOLO

I will not get too technical on how YOLO works, i will focus mostly in the most important parts you need to know about

  • Its very fast which make it suitable for real time applications
  • YOLO processes the image in a single pass through a neural network to make detection and classification
  • Bounding box predictions include coordinates of box center along with width and height of the box, plus the confidence score of the prediction
  • Class prediction, also includes label of the detected objects of the image
  • Multiple scales, it allows detects objects of various scales
  • Pre-trained, it comes with many pre-trained models that have been trained in large datasets
  • Customization, you can train new models for YOLO

Installing YOLO

Create the following file as requirements.txt

cvzone==1.5.6 
ultralytics==8.0.26 
hydra-core>=1.2.0 
matplotlib>=3.2.2 
numpy>=1.18.5 
opencv-python==4.6.0.66 
Pillow>=7.1.2 
PyYAML>=5.3.1 
requests>=2.23.0 
scipy>=1.4.1 
torch>=1.7.0 
torchvision>=0.8.1 
tqdm>=4.64.0 
filterpy==1.4.5 
scikit-image==0.19.3 
lap==0.4.0

Install the libraries with pip3 install -r requirements.txt , Please note that there might be inconsistencies between the package versions and the Python versions, those packages working with Python 3.10

The code

Create the following file as simple_detect.py

#!/usr/bin/env python3 
from ultralytics import YOLO 
import sys 
import cv2 
 
if __name__ == '__main__': 
     
    model   = YOLO('yolov8n.pt') 
    results = model(sys.argv[1],show=True) 
    cv2.waitKey(0) 
    cv2.destroyAllWindows()

The following lines imports the YOLO library and load the pretrained model

from ultralytics import YOLO 
model = YOLO('yolo8n.pt')

This line does the actual processing of an image given from the command line

results = model(sys.argv[1],show = True)

Testing

Lets test the script with this image, some birds at the zoo!

Running the script

simple_detect.py birds.jph

We can see that detected some birds, other birds havent detected at all, we also can see the classification and the confidence level of the classification

Lets try with another one image, me with some cats

Running the script vs this image produced better results but still the stars of the photo (the cats have not been detected)

Conclusion

Computer vision is a new to me concept but is very exciting! i think its not very easy to understand in depth but still the libraries are very useful and without any deep tunning can produce very good results! i hope you enjoyed this article as much i enjoyed writing it.

In Plain English 🚀

Thank you for being a part of the In Plain English community! Before you go: