我不知道扫雷舰是什么，但从我的理解，你想知道哪些瓷砖是蓝色的，然后对你的numpy数组进行必要的更改（如果是别的东西，请纠正我，我会编辑或删除答案）。
以下是将四个图块变为蓝色后的输出：

我做了什么
首先，我阈值的蓝色，发现轮廓的范围约为您的瓷砖。
找到他们的中心看看他们遵循什么模式-

它们都被大约55-57个像素（x，y坐标）分开。其余的很简单。

for i in range(len(coords)):
np_arr[int(coords[i][1]/57)][int(coords[i][0]/57)]=0

蓝色图块的坐标存储在坐标中，np_arr是我的数组。

下面是我使用的代码。最后我不得不重新调整我的数据框架，但只是为了更好的人类阅读。我认为我的代码中有很多可能的改进，但我很高兴我的代码的行为足以满足我的需要。

import cv2
import numpy as np
import pandas as pd

img = cv2.imread('full_snap.png')
"""
some of my blue pixels
R: between 98 and 128
G: between 176 and 211
B: between 255

h: between 210 and 200
s: between 100 and 48
v: between 68 and 100

hsv/2 in opencv and opencv uses BGR not RGB
"""

blue_MIN = np.array([255, 176, 98])
blue_MAX = np.array([255, 211, 128])
"""
https://pythonprogramming.net/color-filter-python-opencv-tutorial/
"""

# find the blue pixels and save it in frame_threshed
frame_threshed = cv2.inRange(img, blue_MIN, blue_MAX)

# find contours in the thresholded image
cnts = cv2.findContours(frame_threshed.copy(), cv2.RETR_EXTERNAL,
    cv2.CHAIN_APPROX_SIMPLE)
cnts = cnts[0] if imutils.is_cv2() else cnts[1]

#print("cnts", cnts)
i = len(cnts)
print("i", i)

# we know we're gonna have x rows of data, where x is the product of
# board_width * board_height
numberOfRows = 81

# check if the length of cnts are euqal to the number of rows/ number of tiles
# then go further

# TODO

# create x,y data
# df = pd.DataFrame(index=np.arange(numberOfRows, 0), columns=('x', 'y'))
d = []
print("d", d)
print(type(d))

for c in cnts:
    # compute the center of the contour
    M = cv2.moments(c)
    cX = int(M["m10"] / M["m00"])
    cY = int(M["m01"] / M["m00"])

    # fill the data with the coords
    d.append({'tilenumber': i, 'X-Value': cX, 'Y-Value': cY})

    # decrease i to go backwards, because tile number 81 is the first contour
    i-=1
    # draw the center of the shape on the image
    cv2.circle(img, (cX, cY), 1, (255, 255, 255), -1)

df = pd.DataFrame(d)

# only for debugging
print("x,y dataframe", df)
cv2.imshow("Image_with_contours",img)

# Destroys all of the HighGUI windows.
cv2.destroyAllWindows()

谢谢大家的帮助！Robinarthur