import cv2.cv2 as cv
import numpy as np


'''
    图片5的文字太清晰了，所以先提取文字，然后多次膨胀，找到文字所在的区域，在Canny边缘检测得到的图像上
    将文字所在区域全置为0（黑色）， 这样就有效的消除了文字对图像的影响

'''


def hough2les(rho, theta, w):
    a = np.cos(theta)
    b = np.sin(theta)
    x0 = a * rho
    y0 = b * rho
    x1 = int(x0 + w * (-b))
    y1 = int(y0 + w * a)
    x2 = int(x0 - w * (-b))
    y2 = int(y0 - w * a)

    k = float(y2 - y1) / float(x2 - x1)
    b = -1*x1*k + y1
    return k, b



src = cv.imread('./Photo/5.jpg', cv.IMREAD_COLOR)
high, width, channel = src.shape
hs_img = cv.cvtColor(src, cv.COLOR_RGB2GRAY)
# cv.imshow('hs', hs_img)
# cv.waitKey(0)

# 高斯滤波
Gaus_img = cv.GaussianBlur(hs_img, (7, 7), 1.5)
# cv.imshow('gauss', hs_img)
# cv.waitKey(0)

# Canny边缘检测
mask_img = cv.Canny(Gaus_img, 130, 142)
cv.imshow('edge', mask_img)
cv.waitKey(0)

elememt_1 = cv.getStructuringElement(cv.MORPH_RECT, (3, 3))

dilation_img = cv.dilate(mask_img, elememt_1, iterations=10)

cv.imshow('di_img', dilation_img)
cv.waitKey(0)

edge_img = cv.Canny(hs_img, 130, 140)
cv.imshow('edge_img', edge_img)
cv.waitKey(0)


for i in range(high):
    for j in range(width):
        if dilation_img[i][j] == 255:
            edge_img[i][j] = 0

cv.imshow('edge_1_img', edge_img)
cv.waitKey(0)

# 霍夫变换
# p 的间隔为0.45 角度间隔为 1度， 长度大于100才被认为是直线
lines = cv.HoughLines(edge_img, 0.45, np.pi / 180, 100)
print(lines)

# result_line_array = []
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 + width * (-b))
        y1 = int(y0 + width * a)
        x2 = int(x0 - width * (-b))
        y2 = int(y0 - width * a)

        cv.line(src, (x1, y1), (x2, y2), (0, 0, 255), 2)
cv.imshow('result_line', src)
cv.waitKey(0)
# 检测交点
'''
    Hough 变换得到多条曲线
    将 角度相差小于0.1 的直线求平均，认为是同一条直线
'''
# 确定两条直线
flag_value = 1
camp_theta_value = lines[0][0][1]
theta_1_sum = lines[0][0][1]
rho_1_sum = lines[0][0][0]
theta_2_value = 0.0
rho_2_value = 0.0

for i in range(1, len(lines)):
    for rho, theta in lines[i]:
        if np.abs(theta - camp_theta_value) <= 1:
            flag_value += 1
            theta_1_sum += theta
            rho_1_sum += rho
        else:
            theta_2_value = theta
            rho_2_value = rho

# 求平均
theta_1_value = theta_1_sum / flag_value
rho_1_value = rho_1_sum / flag_value

line_1_k, line_1_b = hough2les(rho_1_value, theta_1_value, width)
line_2_k, line_2_b = hough2les(rho_2_value, theta_2_value, width)

point_x = int((line_2_b-line_1_b)/(line_1_k-line_2_k))
point_y = int(line_1_k*point_x + line_1_b)


print('{}, {}'.format(point_x, point_x))
cv.circle(src, (point_x, point_y), 10, (0, 238, 0))

cv.imshow('result', src)
cv.waitKey(0)

exit()