road_marking_detection

import cv2
import numpy as np
import matplotlib.image as mpimg
import matplotlib.pyplot as plt

# 칼라이미지.
image_color = cv2.imread('data2/image.jpg')
# cv2.imshow("ori",image_color)

# 우리가 필요한건 그레이 스케일
image_gray = cv2.cvtColor(image_color, cv2.COLOR_BGR2GRAY)
# cv2.imshow("gray", image_gray)


image_copy = image_gray.copy()

# 값이 195 미만인 것들은, 0으로 셋팅.
# print(image_color.shape)
# print(image_copy.shape)

# print(image_copy[ : , :  ] < 195)

image_copy[  image_copy[ : , :  ] < 195  ] = 0

# cv2.imshow('copy', image_copy)


image = cv2.imread('data2/test_image.jpg')
# print('Height = ', int(image.shape[0]), 'pixels')
# print('Width = ', int(image.shape[1]), 'pixels')

# cv2.imshow('Self Driving Car!', image)

gray_img = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
# cv2.imshow('SDC Gray', gray_img)

# HSV 칼라 스페이스로 변경.
hsv_image = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)
# cv2.imshow('HSV', hsv_image)

# Hue 채널만 표시!
H, S, V = cv2.split(hsv_image)
# H = hsv_image[ : , : , 0 ]

# cv2.imshow("Hue", H)

sharp_kernel_1 = np.array( [     
    [0, -1, 0],
    [-1, 5, -1],
    [0, -1, 0]
] )

sharpened_img_1 = cv2.filter2D(gray_img, -1, sharp_kernel_1)

# cv2.imshow('Gray', gray_img)
# cv2.imshow('Sharpen', sharpened_img_1)


sharp_kernel_2 = np.array( [     
    [0, -1, 0],
    [-1, 9, -1],
    [0, -1, 0]
] )
sharpened_img_2 = cv2.filter2D(gray_img, -1, sharp_kernel_2)

# cv2.imshow("sh 2", sharpened_img_2)


# Blurring
blur_img = cv2.GaussianBlur(gray_img, (11,11), 10)
# cv2.imshow("blur", blur_img)


# Sobel 이용한 edge detection
x_sobel = cv2.Sobel(gray_img, cv2.CV_64F, 0, 1, ksize = 7)
# cv2.imshow("sobel x", x_sobel)

y_sobel = cv2.Sobel(gray_img, cv2.CV_64F, 1, 0, ksize = 7)
# cv2.imshow("sobel y", y_sobel)

# 라플라시안은 한번에 수직수평 다 잡는다.
laplacian = cv2.Laplacian(gray_img, cv2.CV_64F)
# cv2.imshow("lapl", laplacian)

# Canny edge detection
threshold_1 = 120
threshold_2 = 200

canny_img = cv2.Canny(gray_img, threshold_1, threshold_2)
# cv2.imshow("canny", canny_img)

# transformation
image = cv2.imread('data2/test_image2.jpg')
# cv2.imshow("ori", image)

# print(image.shape)

M_rotation = \
cv2.getRotationMatrix2D( (image.shape[1]/2,image.shape[0]/2), 90, 0.5 )

rotated_img = cv2.warpAffine(image, M_rotation, (image.shape[1], image.shape[0]))

# cv2.imshow('rotated', rotated_img)

image = cv2.imread('data2/test_image3.jpg')
# cv2.imshow('ori', image)

height = image.shape[0]
width = image.shape[1]

T_matrix = np.array( [
    [1, 0, 120],
    [0, 1, -150]
 ] , dtype = 'float32' )

# print(T_matrix)

translation_image = cv2.warpAffine(image, T_matrix, (width, height))
# cv2.imshow("tran", translation_image)

# Resizing 
resized_image = cv2.resize(image, None, fx=0.5, fy=0.5, interpolation=cv2.INTER_CUBIC)
# cv2.imshow('resize', resized_image)


#region of interest masking
#RoI: 관심영역

image_color = cv2.imread('data2/test5.jpg')
image_gray = cv2.cvtColor(image_color, cv2.COLOR_BGR2GRAY)
# print(image_gray.shape)

# cv2.imshow('d',image_gray)

#zeros 함수는 파라미터로 몇행, 몇열로 만들지 넣어줘야 한다.  
# balack =np.zeros((image_gray.shape[0], image_gray.shape[1]))
# Zeroslike는 행렬안써줘도 된다.
blank = np.zeros_like(image_gray)

ROI = np.array( [[(0,400), (300,250), (450,300), (640,426)]], dtype = np.int32 )
mask = cv2.fillPoly(blank, ROI, 255)
print(mask)


#비트와이즈 연산자 = 메모리의 8bit 끼리 and연산을 한다. 서로 같으면 true(1),
#다르면 false(0). 
masked_image = cv2.bitwise_and(image_gray, mask)
# cv2.imshow('d', masked_image)


#hough transform
image_c = cv2.imread('data2/calendar.jpg')
image_g = cv2.cvtColor(image_c, cv2.COLOR_BGR2GRAY)

image_canny = cv2.Canny(image_g, 50,200,apertureSize=3)



lines = cv2.HoughLines(image_canny,1,np.pi/180,250)

for i in range(len(lines)):
    for rho, theta in lines[i]:
        a = np.cos(theta)
        b = np.sin(theta)
        x0 = a*rho
        y0 = b*rho
        x1 = int(x0 + 1000*(-b))
        y1 = int(y0+1000*(a))
        x2 = int(x0 - 1000*(-b))
        y2 = int(y0 -1000*(a))

        cv2.line(image_c,(x1,y1),(x2,y2),(255,0,0),2)
cv2.imshow('Canny',image_c)


cv2.waitKey(0)
cv2.destroyAllWindows()