image_converter

Image array inspection and conversion utilities.

robotblockset.cameras.image_converter.is_image_array(image: object) → typing_extensions.TypeGuard.<class 'numpy.ndarray'>[source]

Check whether an object looks like an image array.

checks if an object is a numpy array with 3 dimensions, which is the only thing all image formats have in common

robotblockset.cameras.image_converter.is_float_image_array(image: object) → bool[source]: checks if an object is a valid float image array by checking - if it is a valid image array - if it contains floats - and if the first element is in the right range

robotblockset.cameras.image_converter.is_int_image_array(image: ndarray) → bool[source]: checks if an object is a valid int image array by checking - if it is a valid image array - if it contains ints - and if the first element is in the right range

class robotblockset.cameras.image_converter.ImageConverter(image_in_numpy_float_format: ndarray)[source]

Bases: object

Utility class to convert between numpy arrays of different image formats.

Only supports cpu-located images. Convert cuda images to cpu images (if you can afford it) or re-implement with torch.

Note that these conversions may not be optimal, because we use an intermediate numpy float format. So, there may be a conversion from type A to type B that is faster if you do it directly, but we don’t implement that here to keep implementation complexity low. See also https://github.com/airo-ugent/airo-mono/issues/132.

classmethod from_numpy_format(image: ndarray) → ImageConverter[source]

classmethod from_numpy_int_format(image: ndarray) → ImageConverter[source]

classmethod from_opencv_format(image: ndarray) → ImageConverter[source]

classmethod from_torch_format(image: ndarray) → ImageConverter[source]

property image_in_numpy_format: ndarray

property image_in_opencv_format: ndarray

property image_in_torch_format: ndarray

property image_in_numpy_int_format: ndarray

Functions

`is_float_image_array`(image)	checks if an object is a valid float image array by checking - if it is a valid image array - if it contains floats - and if the first element is in the right range
`is_image_array`(image)	Check whether an object looks like an image array.
`is_int_image_array`(image)	checks if an object is a valid int image array by checking - if it is a valid image array - if it contains ints - and if the first element is in the right range

Classes

ImageConverter(image_in_numpy_float_format)

Utility class to convert between numpy arrays of different image formats.