# 硬幣找輪廓
# 橢圓+灰階

import cv2
import numpy as np
from numpy import array, uint8


def findcontour(img: np.ndarray):
    cv2.threshold(img, dst=img, *(89, 255, 0))  #(圖片,參數為最小門檻值,參數最大門檻值).threshold只能灰階 #檢測輪廓

    cv2.morphologyEx(img, dst=img, *(2, array([[0, 0, 0, 1, 0, 0, 0],
                                               [0, 0, 0, 1, 0, 0, 0],
                                               [0, 0, 0, 1, 0, 0, 0],
                                               [1, 1, 1, 1, 1, 1, 1],
                                               [0, 0, 0, 1, 0, 0, 0],
                                               [0, 0, 0, 1, 0, 0, 0],
                                               [0, 0, 0, 1, 0, 0, 0]], dtype=uint8)), iterations=6)

    cv2.erode(img, dst=img, *(array([[0, 0, 1, 0, 0],
                                     [0, 0, 1, 0, 0],
                                     [1, 1, 1, 1, 1],
                                     [0, 0, 1, 0, 0],
                                     [0, 0, 1, 0, 0]], dtype=uint8),), iterations=1)  #腐蝕

    cv2.dilate(img, dst=img, *(array([[0, 0, 0, 1, 0, 0, 0],
                                      [0, 0, 0, 1, 0, 0, 0],
                                      [0, 0, 0, 1, 0, 0, 0],
                                      [1, 1, 1, 1, 1, 1, 1],
                                      [0, 0, 0, 1, 0, 0, 0],
                                      [0, 0, 0, 1, 0, 0, 0],
                                      [0, 0, 0, 1, 0, 0, 0]], dtype=uint8),), **{'iterations': 8})  #膨脹
    return img




if __name__ == '__main__':

    img = cv2.imread("n2.jpg")
    img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)  #圖像從RGB轉到HSV夜色空間(HSV 表示hue、saturation、value)
    #img_gray1 = cv2.cvtColor(img, cv2.COLOR_BGR2HSV) 
    blurred = cv2.GaussianBlur(img_gray, (11, 11), 0)
    binaryIMG = cv2.Canny(blurred, 20, 160)
    
    #cv2.imwrite("9hsvgg_10.jpg",img_gray)
    img_copy = img_gray.copy()
   
    process = findcontour(img_gray)
    process = cv2.Canny(process, 200, 255, apertureSize=5)  #邊緣檢測
    if "3.0" < cv2.__version__ < "3.5":
        _, cnts, hierarchy = cv2.findContours(process, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE) #cv2.RETR_EXTERNAL只取外層的Contour。
    else:
        cnts, hierarchy = cv2.findContours(process, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)
   # cv2.drawContours(img, cnts, -1, (0, 255, 0), 3) #綠邊界  #繪製所有輪廓-1,其餘參數是顏色，厚度ETC　　#cnts為所有Contours
   
      
    for c in cnts:
        if cv2.contourArea(c) < 600: #輪廓面積
          continue
        M = cv2.moments(c)
        Cx = int(M["m10"] / M["m00"])   #M['m00']表示輪廓面積
        Cy = int(M["m01"] / M["m00"])
        cv2.circle(img, (Cx, Cy), 10, (1, 227, 254), -1)  #最小外接圓
        ellipse = cv2.fitEllipse(c) #橢圓
        cv2.ellipse(img, ellipse, (0, 255, 0), 10)  #橢圓
      
        mask = np.zeros(img_gray.shape, dtype="uint8")  #依Contours圖形建立mask
        #mask = np.zeros(img_gray1.shape, dtype="uint8")  #依Contours圖形建立mask
        #mask = cv2.inRange(img_gray1,(224, 16, 55), (25, 255, 255)) 
                  
        
        cv2.drawContours(mask, [c], -1, 255, -1)    # 255→白色, -1→塗滿
        # show the images
        #cv2.imshow("1hsv_10.jpg", img) #顯示原圖(有加線)
       # cv2.imwrite("1_HSV.jpg",img_gray1)
        cv2.imwrite("2hsvg_11.jpg",img)
        cv2.imwrite("2hsvg_m_11.jpg",mask)
        #cv2.imshow("Image1 + Mask", cv2.bitwise_and(img, img, mask=mask))  #顯示疊起來的圖
        cv2.imwrite("2hsvg_bitwise_11.jpg", cv2.bitwise_and(img, img, mask=mask))  # 寫入圖檔

    
    # 按下任意鍵則關閉所有視窗
    cv2.waitKey(0)
    cv2.destroyAllWindows()