如何在opencv中正确select板的轮廓？

Question

我正在尝试确定棋盘的边界（它们以黑条为界）

我执行以下操作

    img = cv2.imread(filename)
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)
    cv2.imwrite("gray.jpg", gray)

    thresh = cv2.inRange(gray, hsv_min, hsv_max)

    kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (7, 7))
    thresh = cv2.dilate(thresh, kernel, iterations = 1)
    thresh = cv2.erode(thresh, None, iterations = 4)

在

之后，我无法获得更好的结果

contours0, hierarchy = cv2.findContours(thresh.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
    for cnt in contours0:
        rect = cv2.minAreaRect(cnt)
        box = cv2.boxPoints(rect)
        box = np.int0(box) 
        cv2.drawContours(img, [box], -1, (255, 0, 0), 0)

    cv2.imshow('contours', img)

有人可以帮助我吗？

Answer 1

这是在 Python OpenCV 中执行此操作的一种方法。基本上，阈值图像。应用一些形态学。获取外部轮廓。然后在黑色背景上画出周长超过某个阈值的最大的。重要的是你得到了 4 个角。然后得到凸包并将其绘制为边界。

输入：

import cv2
import numpy as np

# read image
img = cv2.imread('chess.jpg')

# convert to grayscale
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)

# threshold to binary
thresh = cv2.threshold(gray, 25, 255, cv2.THRESH_BINARY)[1]

# apply morphology
kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (9,9))
morph = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, kernel)
kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (3,1))
morph = cv2.morphologyEx(morph, cv2.MORPH_CLOSE, kernel)

# invert morph
morph = 255 - morph

# get external contours
cnts = cv2.findContours(morph, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
cnts = cnts[0] if len(cnts) == 2 else cnts[1]

# draw white contour on black background
cntr_img = np.zeros_like(morph)
for c in cnts:
    perimeter = cv2.arcLength(c, True)
    if perimeter > 200: 
        cv2.drawContours(cntr_img, [c], 0, 255, 1)

# get all non-zero points
points = np.column_stack(np.where(cntr_img.transpose() > 0))
hull = cv2.convexHull(points)

# draw convex hull vertices on input image
result = img.copy()
cv2.polylines(result, [hull], True, (0,0,255), 2)

# save results
cv2.imwrite('chess_threshold.jpg', thresh)
cv2.imwrite('chess_morph.jpg', morph)
cv2.imwrite('chess_contour.jpg', cntr_img)
cv2.imwrite('chess_convexhull.jpg', result)

# show results
cv2.imshow('thresh', thresh)
cv2.imshow('morph', morph)
cv2.imshow('cntr_img', cntr_img)
cv2.imshow('result', result)
cv2.waitKey(0)
cv2.destroyAllWindows()

阈值图像：

形态图像：

轮廓图：

在输入的副本上产生凸包：

如何在opencv中正确select板的轮廓？

how to correctly select the outline of the board in opencv?

python

opencv

chess

computer-vision