Using string to define an array of numpy arrays
I have an array of images that is X shaped Y 2048x2088. The x-axis has two 20 pixel areas, one at the beginning and one at the end, which are used to calibrate the main area of ββthe image. To access these areas, I can slice the array like this:
prescan_area = img[:, :20]
data_area = img[:, 20:2068]
overscan_area = img[:, 2068:]
My question is how to define these areas in the config file in order to generalize this slice to other cameras, which may have different prescan and jump areas and therefore a different slice is required.
Ideally, something like the lines below would allow a simple view in a specific camera config file, but I'm not sure how to translate those lines into array slices.
prescan_area_def = "[:, :20]"
image_area_def = "[:, 20:2068]"
overscan_area_def = "[:, 2068:]"
Maybe there is something obvious that I am missing?
Thank!
source to share
You can parse the string and use slice
. The following generator expression inside tuple
will create slice objects for you:
tuple(slice(*(int(i) if i else None for i in part.strip().split(':'))) for part in prescan_area_def.strip('[]').split(','))
Demo:
In [5]: import numpy as np
In [6]:
In [6]: a = np.arange(20).reshape(4, 5)
In [7]: a
Out[7]:
array([[ 0, 1, 2, 3, 4],
[ 5, 6, 7, 8, 9],
[10, 11, 12, 13, 14],
[15, 16, 17, 18, 19]])
In [8]:
In [8]: prescan_area_def = "[:, :3]"
In [9]: a[:, :3]
Out[9]:
array([[ 0, 1, 2],
[ 5, 6, 7],
[10, 11, 12],
[15, 16, 17]])
In [10]: indices = tuple(slice(*(int(i) if i else None for i in part.strip().split(':'))) for part in prescan_area_def.strip('[]').split(','))
In [11]: indices
Out[11]: (slice(None, None, None), slice(None, 3, None))
In [12]: a[indices]
Out[12]:
array([[ 0, 1, 2],
[ 5, 6, 7],
[10, 11, 12],
[15, 16, 17]])
source to share
Here's a regular expression approach.
import numpy as np
import re
def slice_from_str(img, s):
REGEX = r'\[(\d*):(\d*), (\d*):(\d*)\]'
m = re.findall(REGEX,s)
if m:
# convert empty strings to None
groups = [None if x=='' else int(x) for x in m[0]]
start_x, end_x, start_y, end_y = groups
x_slice = slice(start_x, end_x)
y_slice = slice(start_y, end_y)
return img[x_slice,y_slice]
return []
img = np.random.rand(2048,2088)
prescan_area_def = '[:, :20]'
image_area_def = "[:, 20:2068]"
overscan_area_def = "[:, 2068:0]"
slice_from_str(img, prescan_area_def)
slice_from_str(img, image_area_def)
slice_from_str(img, overscan_area_def)
source to share