How to handle Data, and Images(11) OpenCV Contours

Lesson Notes in .ipynb file

How to handle Data, and Images(11) - OpenCV Contours

Topics

• OpenCV Contours
• Summary

OpenCV Contours

Input image must be Gray Scaled beforehand

cv2.findContours(image, mode, method): finding Contours from image returns (contours, hierarchy)

• mode: Contour retrieval mode
  • RETER_EXTERNAL: it returns only extreme outer flags.
  • RETER_LIST: It simply retrieves all the contours, but doesn’t create any parent-child relationship.
  • RETER_TREE: It retrieves all the contours and creates a full family hierarchy list.
• method: Contour approximation method
  • CHAIN_APPROX_NONE: stores absolutely all the contour points. That is, any 2 subsequent points $(x1,y1)$ and $(x2,y2)$ of the contour will be either horizontal, vertical or diagonal neighbors, that is, $max(abs(x1-x2),abs(y2-y1))==1$.
  • CHAIN_APPROX_SIMPLE: compresses horizontal, vertical, and diagonal segments and leaves only their end points. For example, an up-right rectangular contour is encoded with 4 points.

cv2.drawContours(image, contour_index, color, thickness): Draws contours outlines or filled contours.

• contour_index: Parameter indicating a contour to draw. (all contours: -1)

import cv2
import matplotlib.pyplot as plt

# image1 is for contours_external, image2 is for contours_tree
image1 = cv2.imread('chanel.jpg')
image2 = cv2.imread('chanel.jpg')

# gray scale the image
image_gray = cv2.cvtColor(image1, cv2.COLOR_BGR2GRAY)
# thresholding
ret, thresh1 = cv2.threshold(image_gray, 127, 255, 0)
ret, thresh2 = cv2.threshold(image_gray, 127, 255, 0)

plt.imshow(cv2.cvtColor(thresh1, cv2.COLOR_GRAY2RGB))
plt.show()

# have to return index 0, because cv2.findContours returns (contours, hierarchy)
contours_external = cv2.findContours(thresh1, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[0]
contours_tree = cv2.findContours(thresh2, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)[0]

# show all contours of the image (we can select by adjusting the value '-1' to 0 to n)
external_image = cv2.drawContours(image1, contours_external, -1, (0, 255, 0), 4)
tree_image = cv2.drawContours(image2, contours_tree, -1, (0, 255, 0), 4)

plt.imshow(cv2.cvtColor(external_image, cv2.COLOR_BGR2RGB))
plt.show()

plt.imshow(cv2.cvtColor(tree_image, cv2.COLOR_BGR2RGB))
plt.show()

Output:

Image After Threshing	CONTOURS_EXTERNAL	CONTOURS_TREE

Summary

• cv2.findContours(): finding Contours from image returns (contours, hierarchy)
  • mode: Contour retrieval mode
    • RETER_EXTERNAL: it returns only extreme outer flags.
    • RETER_LIST: It simply retrieves all the contours, but doesn’t create any parent-child relationship.
    • RETER_TREE: It retrieves all the contours and creates a full family hierarchy list.
  • method: Contour approximation method
    • CHAIN_APPROX_NONE: stores absolutely all the contour points. That is, any 2 subsequent points $(x1,y1)$ and $(x2,y2)$ of the contour will be either horizontal, vertical or diagonal neighbors, that is, $max(abs(x1-x2),abs(y2-y1))==1$.
    • CHAIN_APPROX_SIMPLE: compresses horizontal, vertical, and diagonal segments and leaves only their end points. For example, an up-right rectangular contour is encoded with 4 points.
• cv2.drawContours(): Draws contours outlines or filled contours.