Gesture recognition method and related apparatus

A gesture recognition apparatus comprising:

• a memory, coupled to at least one processor and storing programming instructions for execution by the at least one processor;
• and
• the at least one processor, configured to execute the programming instructions to facilitate the gesture recognition apparatus to perform the following operations:
• obtaining an original point cloud data set, wherein the original point cloud data set includes a plurality of pieces of point cloud data, the plurality of pieces of point cloud data correspond to a plurality of reflected signals, and the plurality of reflected signals correspond to radar signals transmitted by a radar apparatus;
• and
• filtering the original point cloud data set to obtain a first point cloud data set, wherein the first point cloud data set comprises a plurality of frames of first point cloud data subsets,
• wherein each frame of first point cloud data subset corresponds to one frame in the first point cloud data set,
• wherein the plurality of first point cloud data subsets are continuous in time, and each first point cloud data subset comprises a first cluster center, the first cluster center is a cluster center of a plurality of pieces of point cloud data in the first point cloud data subset,
• wherein a maximum horizontal distance between any two first cluster centers meets a first preset condition, and a duration of the first point cloud data set meets the first preset condition, and
• performing gesture recognition based on a feature value of point cloud data in the first point cloud data set, including:
• normalizing the first point cloud data set based on a quantity of frames in the first point cloud data set being inconsistent with a preset quantity of frames, including:
• selecting, based on a time sequence of the first point cloud data set, a union set of point cloud data in a plurality of frames of first point cloud data subsets in the first point cloud data set before normalization, as point cloud data of one first point cloud data subset in the first point cloud data set after normalization,
• wherein point cloud data in a second frame in the first point cloud data set after normalization is composed of a union set of point cloud data in a fifth frame and point cloud data in a sixth frame in the first point cloud data set before normalization,
• wherein the second frame is a second one of the first point cloud data subsets after normalization,
• wherein the fifth frame is a fifth one of the first point cloud data subsets before normalization,
• wherein the sixth frame is a sixth one of the first point cloud data subsets before normalization,
• wherein the feature value comprises at least one of a coordinate value, a distance, or a velocity of point cloud data in a spatial coordinate system;
• and
• wherein based on a relationship between the quantity of frames in the first point cloud data set before normalization and the preset quantity of frames being not an integer-multiple relationship, odd first point cloud data subsets in the first point cloud data set before normalization are selected as the first point cloud data set after normalization, and a last first point cloud data subset and a penultimate first point cloud data subset in the first point cloud data set before normalization are respectively used as a last first point cloud data subset and a penultimate first point cloud data subset in the first point cloud data set after normalization.

A non-transitory computer-readable storage medium comprising computer-executable instructions that are executed by a processor of an apparatus and cause the apparatus to:

• obtain an original point cloud data set, wherein the original point cloud data set comprises a plurality of pieces of point cloud data, the plurality of pieces of point cloud data correspond to a plurality of reflected signals, and the plurality of reflected signals correspond to radar signals transmitted by a radar apparatus;
• and
• filter the original point cloud data set to obtain a first point cloud data set, wherein the first point cloud data set comprises a plurality of frames of first point cloud data subsets,
• wherein each first point cloud data subset comprises a first cluster center, the first cluster center is a cluster center of a plurality of pieces of point cloud data in the first point cloud data subset,
• wherein a maximum horizontal distance between any two first cluster centers meets a first preset condition, and a duration of the first point cloud data set meets the first preset condition;
• and
• perform gesture recognition based on a feature value of point cloud data in the first point cloud data set, including:
• normalizing the first point cloud data set based on a quantity of frames in the first point cloud data set being inconsistent with a preset quantity of frames, including:
• selecting, based on a time sequence of the first point cloud data set, a union set of point cloud data in a plurality of frames of first point cloud data subsets in the first point cloud data set before normalization, as point cloud data of one first point cloud data subset in the first point cloud data set after normalization,
• wherein point cloud data in a second frame in the first point cloud data set after normalization is composed of a union set of point cloud data in a fifth frame and point cloud data in a sixth frame in the first point cloud data set before normalization,
• wherein the second frame is a second one of the first point cloud data subsets after normalization,
• wherein the fifth frame is a fifth one of the first point cloud data subsets before normalization,
• wherein the sixth frame is a sixth one of the first point cloud data subsets before normalization,
• wherein the feature value comprises at least one of a coordinate value, a distance, or a velocity of point cloud data in a spatial coordinate system;
• and
• wherein based on a relationship between the quantity of frames in the first point cloud data set before normalization and the preset quantity of frames being not an integer-multiple relationship, odd first point cloud data subsets in the first point cloud data set before normalization are selected as the first point cloud data set after normalization, and a last first point cloud data subset and a penultimate first point cloud data subset in the first point cloud data set before normalization are respectively used as a last first point cloud data subset and a penultimate first point cloud data subset in the first point cloud data set after normalization.