OpenCV Basics-1

import cv2
import numpy as np

# load the images
img = cv2.imread("img1.jpg",)

# Convert the image into gray scale
imgGray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

# Blur the images using gaussianBlur
# ksize is kernal size, should always in odd, 1 is min and x is max
imgBlur = cv2.GaussianBlur(imgGray, ksize=(15, 15), sigmaX=0)

# Deduct the edge of images, you can increase the threshold to reduce the edge deduction
imgCanny = cv2.Canny(img, threshold1=100, threshold2=100)

# Increase the thickness of edges introduce the dialation, required numpy to use matrix
imgkernal = np.ones((5, 5), np.uint8)
imgDialtion = cv2.dilate(imgCanny, kernel=imgkernal, iterations=1)

# Erode is opposide of dialtion, we decrease the thikness of Canny
imgErode = cv2.erode(imgDialtion, kernel=imgkernal, iterations=1)

# showing the images
cv2.imshow("Color Image", img)
cv2.imshow("Gray Image", imgGray)
cv2.imshow("Blur Image", imgBlur)
cv2.imshow("Edge Deducted Image", imgCanny)
cv2.imshow("Dialated Image", imgDialtion)
cv2.imshow("Erode Image", imgErode)

# show the image for 5 seconds, 0 means indefinite
cv2.waitKey(10000)

# destroy all the image windows
cv2.destroyAllWindows()

import cv2
import numpy as np

#reading the image
img = cv2.imread('Resources/k_cards.jpg')

# set the width and height of perspective
width, height = 250, 350

# pt1 have upper left, upper right, lower left, lower right
pt1 = np.float32([[298, 339], [529, 416], [131, 650], [384, 741]])

# pt2 have the perception we want for the image
pt2 = np.float32([[0, 0], [width, 0], [0, height], [width, height]])

# get the perspective of our cordinates to actual cordinates
matrix = cv2.getPerspectiveTransform(pt1, pt2)

# wrap into output to see the Images into correct perspective
imgOutput = cv2.warpPerspective(img, matrix, (width, height))

# show image
cv2.imshow("Image", img)
cv2.imshow("Changed Image", imgOutput)

# show the image for 5 seconds, 0 means undefinite
cv2.waitKey(10000)

# destroy all the image windows
cv2.destroyAllWindows()

import cv2
import numpy as np

# reading the image
img1 = cv2.imread('Resources/img1.jpg')
img2 = cv2.imread('Resources/img2.jpg')
img3 = cv2.imread('Resources/img3.jpg')

# Horizontal stack
imgHorizontalStack = np.hstack((img1, img3))

# Vertical Stack
imgVerticalStack = np.vstack((img1, img3))

# show image
cv2.imshow("Horizontal Stack Image", imgHorizontalStack)
cv2.imshow("Vertical Stack Image", imgVerticalStack)

# show the image for 5 seconds, 0 means indefinite
cv2.waitKey(10000)

# destroy all the image windows
cv2.destroyAllWindows()

import cv2
import numpy as np

# function to convert the images to match the channel
def stackImages(img_list=[], stack_type=0):

    # Create the empty list to store images
    list_of_img = []
    # Iterate the images
    for img in img_list:
        # check if the image is color or not , color will have (row, column, channel) whereas gray don't have channel

        if len(img.shape) < 3:
            # transform the image to color level
            img = cv2.cvtColor(img, cv2.COLOR_GRAY2RGB)

        # add to list
        list_of_img.append(img)

    # check if we need to stack horizontal or vertical, by default is horizontal
    if stack_type == 0:
        imgStack = np.hstack(tuple(list_of_img))
    else:
        imgStack = np.vstack(tuple(list_of_img))

    # show the images
    cv2.imshow("Stacked Image", imgStack)

    # show the image for 5 seconds, 0 means indefinite
    cv2.waitKey(10000)

    # destroy all the image windows
    cv2.destroyAllWindows()

if __name__ == '__main__':
    # read the images
    img = cv2.imread('img1.jpg')
    img2 = cv2.imread('img2.jpg', 0)
    img3 = cv2.imread('img3.jpg', -1)

    # Call the function
    stackImages([img, img2, img3])

import cv2
import numpy as np

# Reading the image
img = cv2.imread("lamb.jpg")

######################  IMAGE RESIZE #################
# find the size of the images
print(img.shape)
"""
(1333, 2000, 3)  # width, height and channel
"""
# resize the image size
imgResize = cv2.resize(img, (300, 250))

# print the new image size
print(imgResize.shape)
"""
(480, 640, 3)  # width, height and channel
"""
# writing this image into new file
cv2.imwrite("Resources/lamb.jpg", imgResize)

# showing the small image
cv2.imshow("Small Size", imgResize)

######################  IMAGE CROPPING #################

ImgCropped = imgResize[0:img.shape[0], 100: img.shape[1]]  # [height, width]
cv2.imshow("Cropped Size", ImgCropped)

# waiting for the key or 10 secs
cv2.waitKey(10000)

# Destroy all the windows
cv2.destroyAllWindows()