Kinect Uygulamaları için Veri Transfer Platformu Tasarımı

Abstract

In recent years, software, hardware, and algorithms have come to fruition. These developments in technology have also affected sensor technologies. The Kinect sensor, initially marketed as a gaming device, has been met with great interest by both researchers and developers. The Kinect sensor has been used in different areas in the literature for different purposes. All data from the Kinect sensor is transmitted to developers through the Software Development Kit ( SDK ) developed by Microsoft. The Kinect sensor produces between 240 and 270 thousand points per second in normal conditions, depending on scene complexity. The purpose of this work is to design a data transfer platform for Kinect applications. The developed platform works on client server architecture. The platform, which has different scenarios for online and offline communication situations, also offers several filtering and encryption algorithms. The platform uses the Point Cloud Library (PCL), a large-scale open source project for 2D / 3D image and point cloud processing. VoxelGrid (VG) Filter, Outlier Filter, Histogram Based Conditional Filter, Octree-based Compression and PGP Encryption methods are also available on request. In addition, a special data structure has been developed for Kinect applications. WebRTC middleware software is used for online communication. Thanks to all these steps, unnecessary data points have been cleaned, compressed, secured and data packages suitable for the developed data structure have been obtained. Because of the filtering, a compression ratio of 19.96% has been obtained. Application-based or client-based filtering is provided through custom design. A file compression result of 10.38% has been obtained with the file compression approach applied after filtering. The presented platform will provide performance for Kinect applications used by researchers and developers.