CN112757302A

CN112757302A - Control method of portable dining-assistant robot

Info

Publication number: CN112757302A
Application number: CN202110012450.2A
Authority: CN
Inventors: 陈殿生; 赵学毅; 袁福; 罗亚哲
Original assignee: Beihang University
Current assignee: Beihang University
Priority date: 2021-01-06
Filing date: 2021-01-06
Publication date: 2021-05-07

Abstract

The invention designs a control method of a portable dining-assistant robot, which is based on a face recognition positioning and stopping triggering mode. The dining assistant robot identifies the mouth position of the user: and (3) measuring distance by adopting a binocular camera imaging and point calibration method. Thereby obtaining the spatial position of the mouth. The dining assistant robot identifies the intention of the user: detecting whether the old man is nodding or not, and starting the mechanical arm when nodding; shaking the head, the execution unit stops working, and interference caused by other conditions such as shaking and the like is eliminated. According to the technology, a low-cost and high-efficiency positioning mode is realized.

Description

Control method of portable dining-assistant robot

Technical Field

The invention relates to the technical field of robots, in particular to the field of robot control, and discloses a control method based on facial recognition and binocular ranging.

Background

The weak self-care ability of the empty nest old people is a problem to be solved urgently in the society at present. About 900 million empty nesters have difficulty eating food.

The dining-assistant robot designed for the old has great market demand. The dining-assistant robot can enhance the self-care ability of the old, reduce the psychological pressure of the old and reduce the burden of children. Considering economic factors, the dining-assistant robot is economical and practical; considering the cultural degree, the dining assistant robot should be easy to operate.

Disclosure of Invention

The invention designs a control method of a portable dining-assistant robot, which is based on a face recognition positioning and stopping triggering mode to solve the economic problem and the operation problem mentioned in the background technology.

A control method of a portable dining-assistant robot is characterized in that the system comprises the following steps: camera group, stop trigger system.

The camera group collects images, recognizes faces, takes the lower half part of the faces as an interested area, and then recognizes the mouth.

Preferably, after the mouth is identified, a binocular distance measurement principle is utilized, firstly, a Zhang Zhengyou calibration method is adopted, and then the position of the mouth is located by combining a human eye parallax principle.

The stopping triggering system judges whether the old people shake the head or not by detecting the moving distance and frequency of the eyes in unit time based on the recognition of the eyes, and when the old people shake the head, the execution unit stops working.

The invention has high identification precision and low hardware cost; the old can choose to eat or not by nodding and shaking the head, and the operation is simple.

Drawings

FIG. 1 is a system architecture diagram of the present invention.

Fig. 2 is a flow chart of binocular positioning.

Detailed Description

The present invention is directed to a novel control method and provides a novel human-computer interaction method to solve the above-mentioned problems in the background art.

The invention provides a technical scheme that a dining assistant robot identifies the intention of a user: detecting the distance and frequency of the movement of the eyes of the person in unit time, and judging whether the old man nods or not, wherein the mechanical arm is started when the old man nods; when the old person shaking the head is detected, the execution unit stops working, and interference caused by other conditions such as shaking is eliminated. To achieve this function, the programming principle is as follows: firstly, the face is calibrated, and the positions of eyes and the positions of eyeballs are calibrated. And then listing a position coordinate matrix and a translation matrix, and judging the moving pose of the eyeball by utilizing matrix transformation. Only parallel movement with the left and right limit distance larger than 4mm is effective data, and the others are judged as interference. Three dishes are arranged in the dinner plate, and the dishes are switched leftwards when the eyeball moves leftwards; when the eyeball moves to the right, dishes are switched to the right.

The invention provides a technical scheme that a dining assistant robot identifies the mouth position of a user: firstly, eliminating distortion of an image collected by a camera, calculating parallax formed by a target point on a left view and a right view, and firstly matching two corresponding image points of the target point on the left view and the right view. To reduce the matching search range, we can use epipolar constraint to reduce the matching of corresponding points from two-dimensional search to one-dimensional search. And then binocular correction is carried out, and the function is to strictly correspond the two images after distortion removal, so that epipolar lines of the two images are exactly on the same horizontal line, thus any point on one image and the corresponding point on the other image have the same line number, and the corresponding point can be matched only by one-dimensional search on the line.

The binocular camera adopts a distance measurement principle similar to human eyes. Human eyes can perceive the distance of an object because the images of the same object presented by the two eyes are different, which is also called as parallax. The farther the object distance is, the smaller the parallax error is; conversely, the greater the parallax. The magnitude of the parallax corresponds to the distance between the object and the eyes, and the spatial position of the mouth is obtained.

The invention realizes the depth information positioning through binocular ranging, and then judges the coordinates of the mouth position by using the parameters built in the camera. The internal parameters of each camera are obtained first, the relative position between the two cameras (namely the translation vector t and the rotation matrix R of the right camera relative to the left camera) is measured through calibration, and the stereoscopic coordinate of the depth position is obtained through the matrix transformation of the projection matrix.

Claims

1. A control method of a portable dining assistant robot is characterized in that: the method is based on face recognition positioning and stopping trigger modes. The system comprises: camera group, stop trigger system.

2. The face recognition localization method of claim 1, wherein the camera group collects images, recognizes faces, and recognizes mouths by using the lower half of the faces as an interested area. After the mouth is identified, the binocular distance measuring principle is utilized, firstly, a Zhang Zhengyou calibration method is adopted, and then the position of the mouth is positioned by combining the human eye parallax principle.

3. The stop trigger system according to claim 1, wherein the system determines whether the elderly is shaking the head by detecting the distance and frequency of movement of the eyes in a unit time based on the recognition of the eyes, and the execution unit stops working when the elderly is detected to shake the head.

4. The method as claimed in claim 1, wherein the face recognition determines whether to switch dishes by detecting the distance and frequency of the movement of the human eyes in a unit time. The function can help patients with quadriplegia and aphasia to select proper dishes.