mindelec.data.PointCloudSamplingConfig

class mindelec.data.PointCloudSamplingConfig(sampling_mode, bbox_type, mode_args=None, bbox_args=None)[source]

Sampling space config for PointCloud-Tensor generation.

Parameters
  • sampling_mode (int) – Point sampling method. 0(UPPERBOUND) and 1(DIMENSIONS) are supported.

  • bbox_type (int) – Bounding box type for sampling space, only supports cube-shape sampling space. 0(STATIC) and 1(DYNAMIC) are supported.

  • mode_args (Union[int, tuple]) – sampling upperbound number for SamplingMode. Default: None

  • bbox_args (tuple) – bounding_box arguments for sampling, has different definition in different bbox_type. Default: None.

Raises
  • TypeError – if sampling_mode is not an int.

  • TypeError – if bbox_type is not an int.

  • TypeError – if mode_args is not one of int or tuple.

  • TypeError – if bbox_args is not a tuple.

  • TypeError – if sampling_mode is 0 but mode_args is not int.

  • TypeError – if sampling_mode is 1 but mode_args is not a tuple of three integers.

  • ValueError – if sampling_mode is 1 but the length of mode_args is not three.

  • ValueError – if sampling_mode not in [0(UPPERBOUND), 1(DIMENSIONS)].

  • ValueError – if bbox_type not in [0(STATIC), 1(DYNAMIC)].

Supported Platforms:

Ascend