Research On 3D Indoor Visual Positioning And Orienting System Based On QR Code

During autonomous navigation,the mobile robot is inseparable from own position and orientation in the global environment.Only by accurately grasping position and orientation can better achieve autonomous navigation.The current QR code positioning and orienting method is generally limited to the two-dimensional space.For example,it can only calculate the two-dimensional position(X,Y)and yaw angle,or three-dimensional position(X,Y,Z)but lack of attitude angle,which limits the application scope of the system.Therefore,in order to realize the positioning and orienting in a general three-dimensional space,this dissertation designs a visual positioning and orienting system by QR code,which can calculate the three-dimensional position(X,Y,Z)and three-dimensional orientation(roll angle,pitch angle,yaw angle)of the robot in the global environment.The main research content and innovation of this dissertation include the following:(1)Process the QR code images collected by the camera,including preprocessing,QR code detection,correction and decoding.Aiming at the problem that the traditional QR code detection algorithm based on the proportional feature has a low success rate when the camera is shot at a large tilt angle,this dissertation proposes an algorithm based on the contour inclusion relationship and the center overlap relationship.First,find the three-level contour in the image that conforms to the nested inclusion relationship,and then determine the contour of the location pattern by comparing the distance between the center coordinates of the contour,and realize the detection of the QR code.In addition,before finding the contours that meet the conditions,a contour pre-screening method can effectively eliminate the contours that are obviously too large or too small,which greatly improves the efficiency of QR code detection.Experimental results show that this algorithm can effectively detect the position of the QR code in the image when the camera tilt angle is large,and meet the needs of the robot's moving range.(2)Aiming at the two-dimensional positioning limitations of the existing QR code indoor positioning methods,this dissertation designs a method to achieve three-dimensional positioning and three-dimensional orienting of the robot by QR code.First,define a QR code coordinate system as an intermediary connecting the robot and the global environment.Then use the EPn P +Levenberg-Marquardt algorithm to solve the camera's position and orientation in the QR code coordinate system.Finally combine the position and orientation of the QR code in the global environment stored in the database to realize the camera three-dimensional positioning and three-dimensional orienting in the global environment.Experimental results show that this method can effectively measure the three-dimensional position and three-dimensional orientation of the robot,with high accuracy and good stability,and the results will not diverge with the increase of time or moving distance.The QR code visual positioning and orienting system in this dissertation has high accuracy,good flexibility and low hardware deployment cost requirements.It only needs a monocular camera and QR code to calculate the three-dimensional position and three-dimensional orientation of robot.It can be used as a simple and convenient method for absolute positioning and orienting and be deployed in indoor environments with weak GPS signals or few visual features to provide robots with indispensable three-dimensional position and three-dimensional orientation services,which has great practical value and application prospects.