Knowledge Base

What is Image Processing? How is it Done? A Basic Guide with Examples

Image processing is the process of analyzing, interpreting, and transforming an image. Digital image processing aims to process visual data through computers, performing operations such as information extraction, filtering, enhancement, or automation. This field is used in many sectors such as artificial intelligence, machine learning, robotics, health, and security.

What is Image Processing?

Image processing is divided into two main categories:

1. Analog Image Processing: Operations performed with physical tools such as optical lenses and filters.

2. Digital Image Processing: Processing images numerically (with pixel data) in software.

In this article, we will focus on digital image processing.

Goals in Image Processing

• Noise reduction

• Image enhancement

• Edge detection

• Object recognition

• Segmentation (dividing the image into regions)

• Object tracking

• Image classification (with CNN)

How to Do It? Tools and Libraries Used

Common languages:

• Python

• C++

• MATLAB

Libraries:

• OpenCV (C++ / Python)

• Pillow (Python)

• scikit-image (Python)

• TensorFlow/Keras (Deep learning-supported image processing)

✅ Image Processing Steps

1. Image Input:

import cv2
img = cv2.imread("resim.jpg")

2. Pre-processing:

gri = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
blur = cv2.GaussianBlur(gri, (5, 5), 0)

3. Edge Detection:

kenarlar = cv2.Canny(blur, 50, 150)

4. Segmentation:

_, thresh = cv2.threshold(gri, 127, 255, cv2.THRESH_BINARY)

5. Contour Detection:

contours, _ = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
cv2.drawContours(img, contours, -1, (0,255,0), 2)

Example Projects

1. Face Recognition

With OpenCV's Haar cascade classifier:

face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
faces = face_cascade.detectMultiScale(gri, scaleFactor=1.1, minNeighbors=5)

2. License Plate Recognition

• Segment the license plate region and perform text recognition with OCR (Tesseract).

3. Hand Gesture Control

• Separate the hand region from the background.

• Detect the number of fingers with contour analysis.

4. Noise Cleaning and Filtering

median = cv2.medianBlur(img, 5)

5. Processing Live Video with Camera

cap = cv2.VideoCapture(0)
while True:
    ret, frame = cap.read()
    gri = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    cv2.imshow("Görüntü", gri)
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break
cap.release()
cv2.destroyAllWindows()

Advanced Applications

Resources

• https://opencv.org/

• https://docs.opencv.org/

• https://pyimagesearch.com/

✅ Conclusion

Image processing is a powerful discipline that gives computers the ability to understand visual data. With libraries like OpenCV, many applications can be developed, from simple image filtering to complex object recognition projects. This technology, which is beneficial in every sector that works with visual data, is one of the cornerstones of the world of artificial intelligence and robotics.