如何获取图像内矩形内的像素值

Question

I have an image, where four corner points are defined. Now I want to get the pixel values of the region, that is defined by the 4 corners. Problem is, although it is a rectangle, it has a "slope", which means neither the two upper corner points nor the lower one are at the same height. How can I still solve this issue?

I have not found anything for this yet.. I'd appreciate any kind of support: :)

Answer 1

You can do that with Python/OpenCV by first drawing a white filled polygon on black background as a mask from your four points. The use np.where to locate and then print all the points in the image corresponding to the white pixels in the mask.

Input:

import cv2
import numpy as np

# read image
image = cv2.imread('lena.png')

# create mask with zeros
mask = np.zeros((image.shape), dtype=np.uint8)

# define points (as small diamond shape)
pts = np.array( [[[25,20],[30,25],[25,30],[20,25]]], dtype=np.int32 )
cv2.fillPoly(mask, pts, (255,255,255) )

# get color values
values = image[np.where((mask == (255,255,255)).all(axis=2))]
print(values)

# save mask
cv2.imwrite('diamond_mask.png', mask)

cv2.imshow('image', image)
cv2.imshow('mask', mask)
cv2.waitKey()

Mask:

Results:

 [[108 137 232]
 [104 134 232]
 [108 136 231]
 [106 134 231]
 [109 133 228]
 [108 136 229]
 [109 137 230]
 [110 135 232]
 [103 126 230]
 [112 134 228]
 [114 136 228]
 [111 138 230]
 [110 137 233]
 [103 135 234]
 [103 126 230]
 [101 120 226]
 [108 137 230]
 [112 133 228]
 [114 136 227]
 [115 139 232]
 [112 137 232]
 [105 134 233]
 [102 128 232]
 [ 98 119 226]
 [ 93 105 220]
 [108 139 230]
 [110 137 230]
 [112 135 230]
 [113 135 230]
 [111 138 231]
 [112 139 232]
 [109 134 233]
 [101 128 232]
 [100 120 224]
 [ 90 104 221]
 [ 87  95 211]
 [111 138 229]
 [109 135 231]
 [109 136 230]
 [113 141 233]
 [110 139 233]
 [105 136 234]
 [101 127 232]
 [ 95 117 225]
 [ 90 107 220]
 [110 137 231]
 [110 138 231]
 [107 140 236]
 [110 139 233]
 [104 135 234]
 [105 130 231]
 [ 92 116 227]
 [114 141 234]
 [112 142 235]
 [111 140 235]
 [111 138 234]
 [110 132 232]
 [114 140 234]
 [108 140 233]
 [107 134 233]
 [107 140 235]]

Answer 2

It is not very easy to iterate through a slanted rectangle. Therefore, what you can do is to rotate the whole image such that the rectangle is parallel to the sides again.

For this, you can compute the slope of one side as difference in the y coordinate over the difference in the x coordinate of the corners. The value you get is the slope. The arctangent of the slope is the angle to the horizontal. You need to rotate the image with the opposite (negative) of this value.

To make it more efficient, you can crop a bit the image.