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
47
48
|
Load an image from disk into a numpy array
SYNOPSIS
image = \
mrcal.load_image("scene.jpg",
bits_per_pixel = 8,
channels = 1)
## image is now a numpy array of shape (height,width) containing the
## pixel data
This is a completely uninteresting image-loading routine. It's like any other
image-loading routine out there; use any that you like. This exists because cv2
is very slow.
This wraps the mrcal_image_TYPE_load() functions. At this time I support only
these 3 data formats:
- bits_per_pixel = 8, channels = 1: 8-bit grayscale data
- bits_per_pixel = 16, channels = 1: 16-bit grayscale data
- bits_per_pixel = 24, channels = 3: BGR color data
With the exception of 16-bit grayscale data, the load function will convert the
input image to the requested format. At this time, asking for 16-bit grayscale
data requires that the input image matches that format.
If we ask for an 8-bit image, but pass a 16-bit image file, we will apply
stretch equalization to the input first.
If the bits_per_pixel, channels arguments are omitted or set to <= 0, we will
load the image in whatever format it appears on disk.
ARGUMENTS
- filename: the image on disk to load
- bits_per_pixel: optional integer describing the requested bit depth. Must be 8
or 16 or 24. If omitted or <= 0, we use the bit depth of the image on disk
- channels: optional integer describing the number of channels in the image.
Integer. Must be 1 or 3. If omitted or <= 0, we use the channel count of the
image on disk
RETURNED VALUE
A numpy array containing the pixel data
|
