Research Of 3D Panoramic Electronic Map Based On Depth Camera

With the rapid development of science and technology,artificial intelligence technology is maturing.The development of robots also moves from industrial robots to mobile robots for people's convenience.For the foreseeable future,robots may become an integral part of people's lives section.SLAM means robots are able to create their own maps in a new,unknown environment and base their knowledge on positioning and path planning.SLAM technology has a wide range of applications,in the need to locate,navigate and can't be obtained in advance of the scene map.It is the premise of mobile robot to the intelligent direction.This paper presents a real-time panoramic camera based on depth camera technology solutions to build a platform system.The depth image acquired by Kinect v2.0 in real time is transformed into coordinate of three-dimensional space through the transformation of image coordinate system and camera coordinate system.The point cloud registration algorithm is used to solve the rigid transformation matrix in sensor movement in real time.The depth images of different frames are converted into the same coordinate system.It is displayed for subsequent positioning and path planning.First of all,this paper introduces the hardware and software composition of the electronic map system,introduces the different imaging principles of the depth camera of the hardware device.Then it introduces the software development platform of the system and the main components in the software system,including bilateral filtering,image pyramid and so on.The principle and implementation of the system are described.And an intuitive picture comparison is given.In the core part of the algorithm,this paper introduces the Iterative Closest Point(ICP)algorithm theory in point cloud registration.The algorithm is an iterative algorithm,which includes projection method,point cloud initial registration algorithm,matching point determination algorithm and rigid body transformation matrix calculation algorithm.In addition,some parts of the paper are compared and analyzed by classical methods and improved methods,and the feasibility of this algorithm in this experimental system is demonstrated.Calibration of camera parameters mainly introduces calibration and correction of camera geometry parameters,registration of Kinect depth lens and RGB lens.Calibration of the camera geometric parameters using Zhang Zhengyou calibration method,the paper elaborated the principle of the method and calibration steps,and correct the Kinect depth lens.The Experimental Results Analysis section shows the various parts of the previous system software composition and has carried on the analysis and the summary to each link.