computer_science:computer_vision:opencv:object_identification_tensorflow_opencv_imageai_yolo [Carlos Sousa Wiki]

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== Object Detection with Tensorflow, OpenCV, ImageAI and YOLO ======

===== Requirements: =====

Be sure to have the [[:computer_science:computer_vision:opencv:object_identification_tensorflow_opencv_imageai_yolo_requirements|requirements]] installed.

===== Project Start - Importing Libraries and Testing =====

==== Initialize the libraries ====

<code python>
import cv2 as cv #Loads ComputerVision / OpenCV
from imageai.Detection import ObjectDetection as od # Loads ImageAI

import numpy as np # Imports NumPY
import requests as req # Imports Requests
import os as os # Imports OS Module


</code>

==== Testing if OpenCV is working: ====

<code python>
url = 'https://p7.hiclipart.com/preview/124/937/193/architectural-engineering-engineer.jpg' # Image to be Downloaded
r = req.get(url) # Pass the URL as a Request.Get
with open('testimage.jpg', 'wb') as outfile: # Writes the Image to the system
    outfile.write(r.content)
img = cv.imread('testimage.jpg') # Reads an image with OpenCV
window_name = 'image' # Defines a Window Name
cv.imshow(window_name, img) # Show the Image on a Window
cv.waitKey(0) # Wait for any Key to be pressed
cv.destroyAllWindows() # Destroy (all) previous created Windows


</code>

==== Preparing for the upcoming project: ====

Now that we tested that OpenCV is working let's edit that last piece of code, change it to a function, so we can reutilize it often. The complete code until now should look like this:

<code python>
import cv2 as cv
from imageai.Detection import ObjectDetection as od

import numpy as np
import requests as req
import os as os

def showImage(img):
    window_name = 'image'
    cv.imshow(window_name, img)
    cv.waitKey(0)
    cv.destroyAllWindows()


</code>
=====   =====

===== Downloading Images to train the ML =====

Run this [[:computer_science:computer_vision:opencv:object_identification_tensorflow_opencv_imageai_yolo_download_images_script|code]] to download 2 type of images that we will be using. One being only of people and the other to download images of hats. This images will be contained on they own folder since we will be using that to train the Machine Learning.

===== External References: =====

[[https://www.codeproject.com/Articles/5270236/Installing-OpenCV-and-ImageAI-for-Object-Detection|https://www.codeproject.com/Articles/5270236/Installing-OpenCV-and-ImageAI-for-Object-Detection]]