A task decision method and device of a robot and an electronic device

Abstract

This application relates to the field of robotics, and more particularly to a task decision-making method, apparatus, and electronic device for a robot. The method includes: during the execution of a first task, identifying part information corresponding to multiple parts from an image containing at least one interactive object; determining a first correlation score for different parts based on bounding box information of different parts among the multiple parts, and determining a second correlation score for the different parts based on key point coordinates of the different parts; determining whether the different parts belong to the same interactive object based on a weighted sum of the first and second correlation scores, thereby obtaining an object recognition result for the image; and executing the first task based on the object recognition result of the image. Through the above method, it is possible to determine part information belonging to the same interactive object, thereby forming a continuous cognition and memory of the posture of a complete interactive object during task execution.

Images