1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
|
Finds corner or blob features in an image
SYNOPSIS
corners = mrgingham.find_points(image)
[ corners is an array of shape (N,2), the unordered list of ]
[ detected feature coordinates in the image ]
This function runs just the initial stage of the mrgingham processing:
- pre-process image
- find chessboard corners (or blobs)
This function does NOT look for the grid in the feature set
No broadcasting is supported by this function
ARGUMENTS
- image: numpy array of shape (H,W) and dtype np.uint8. This is the image we're
processing. Must be densely-stored, grayscale image
- blobs: optional boolean, defaulting to False. If True, we look for
black-on-white blobs, instead of chessboard corners. If blobs: we MUST have
image_pyramid_level==0
- image_pyramid_level: optional integer defaulting to -1. This can be given to
operate on a downsampled version of the image. 0 means "original image", 1
means "downsample by a factor of 2 in each dimension", 2 means "downsample by
a factor of 4 in each dimension" and so on. image_pyramid_level < 0 (the
default) means "start with a downsampled image, and then refine the results by
repeatedly reducing the downsampling. If blobs: image_pyramid_level==0 is the
only allowed option
- debug: optional boolean, defaulting to False. If True, this will dump various
intermediate results into /tmp and it will report more stuff on standard out.
Most of the intermediate results are self-plotting data files. Run them. This
does the same thing as "mrgingham --debug"; the Python output is unaffected.
RETURNED VALUE
A numpy array of shape (N,2) containing unordered pixel coordinates of the
discovered features. If no features were found, and array of shape (0,2) is
returned.
|
