Wireless Video Transmission and Reproduction of Indoor Scenes Based on the Real-time Monitoring

Abstract

It is proved that the video wireless transmission and reproduction system for indoor scene real-time monitoring is not nonsense towards implementation and application, for people’s requirements for security, and for the special populations to see the indoor scenes where they can’t reach. But the traditional real-time monitoring system could only show us the scene at the current video frame. To solve this problem, this thesis proposes and implements a method for real-time monitoring and reproducing the map of the area using the 3d point cloud simultaneously.

Kinect has been used as an image acquisition device, which receives color images and the corresponding depth information for every pixel at the same time. This information will be processed in the mobile terminal to acquire the point cloud. The images captured by the Kinect will be transmitted to the principle computer and displayed by the principle computer. And the images as well as the point cloud will be processed in the terminal before transmission. The principle computer will finish the registration of the point cloud, to complete the map generating.

The method mentioned in this paper, using the Kinect to get point cloud, collects information to reconstruct the scene for real-time monitoring. Unlike traditional real-time monitoring systems, some related theories of the computer vision and some methods in the SLAM (Simultaneous Localization and Mapping) of the mobile robot are needed. On the basis of the real-time monitoring of point cloud registration, ICP (Iterative Closest Point) algorithm is used in order to complete the global scenario generation, at the same time, in order to improve the real-time performance of system, the use of Surf on color image features detector to extract and match, based on which point cloud data filtering can be achieved.

Key words: Real-time Monitoring, Map Reconstruction, Kinect, Wireless Transmission, Registration of Point Cloud

摘要

由于人们对于安全的需求，以及特殊人群在移动存在障碍情况下能够实现对整个室内场景监控的需要，应用于室内场景实时监控的视频无线传输与复现系统有其实现以及面向应用的价值。但传统意义上的实时监控系统只能在接收端实时再现当前视频帧的场景，不能对已采集区域有一个全面的把握，针对此问题，本文提出并初步实现了一种在实时监控的同时利用三维点云实现场景复现的方法。

Kinect作为图像采集装置，同时采集到彩色图像及其对应像素点的深度信息，并在移动终端进行处理得到三维场景的点云信息。移动终端在获得三维点云后对彩色图像以及点云进行预处理后，通过无线局域网传输给上位机，再由上位机对接收到的每帧彩色图像进行实时的显示，并对每帧点云信息进行位姿的变换完成点云的配准工作，从而完成三维场景的复现。

本文所实现的内容，是在实时监控的同时，利用Kinect采集到信息得到室内环境点云数据，并在上位机完成三维场景的重建。与传统实时监控系统不同的是，本文中需要运用计算机视觉相关理论以及移动机器人在定位和构图中所涉及的一些方法，在实时监控的基础上运用ICP算法进行点云的配准从而完成全局场景的生成，与此同时为了提高系统的实时性，在彩色图像上利用SURF特征检测算子来进行特征的抽取与匹配，并以此为基础来进行点云数据的筛选工作。

关键词：实时监控，场景复现，Kinect，无线传输，点云配准

Recommended citation: Q. Sun. Wireless Video Transmission and Reproduction of Indoor Scenes Based on the Real-time Monitoring. B.S. thesis, Beihang University, Beijing, China, 2013. http://sunqinxuan.github.io/files/publications-2013-06-01-B.S.Thesis.pdf