Camera angle calibration method during robot speech localization process

Abstract

The invention discloses a camera angle calibration method during a robot speech localization process. The method comprises a first step of speech coarse localization in which the camera is rotated to an approximate speech source direction according to the speech information, a second step of human face detection in which the camera is used for detecting whether a human face exists and a third step is carried out if the camera detects the human face, a third step of human face angle deviation calculation in which a rectangular coordinate axis is built with the camera center as the original point, the coordinates [X, Y] of the human face center are localized, and the coordinates [X, Y] are deviation between the human face center and the camera center, and a fourth step of angle adjustment in which the angle degreeX of the camera rotating along the X axis and the angle degreeY of the camera rotating along the Y axis are converted according to the deviation between the human face coordinates and the camera center and then, the camera is rotated to enable the human face center to be aligned with the camera center according to the degreeX and the degreeY. The camera alignement precision during the speech localization process is improved.

Description

technical field [0001] The invention relates to the technical field of voice positioning, in particular to a camera angle calibration method for a robot voice positioning process. Background technique [0002] With the continuous development of electronic computer science, the research on intelligent robots has become more and more in-depth, and they carry more and more functions. Highly intelligent robots with functions such as vision, hearing, and smell have appeared. System, collect information through the recognition system, and obtain recognition results through the information collected by the processor and its processing system. The common face recognition system uses a camera to collect face images, and the identity of the person can be identified after processing. , but it is not enough for a highly intelligent robot to realize a series of complex operations through a face recognition system, such as a robot that can interact with people, and it also carries a 360° ...

AI Technical Summary

Problems solved by technology

[0003] However, there is still a deviation in the voice positioning technology of the robot. In the above voice positioning process, the voice information can only provide the robot with a rough orientation, and it is difficult to accurately find the center of the face after turning the camera (that is, turning the robot head). The location may cause the camera to acquire images that are not clear enough, and there may be recognition errors. In operation, the user can only find the point that is aligned with the camera. Therefore, the voice positioning technology of the robot needs to be improved.

Images