Book Image

OpenNI Cookbook

By : Soroush Falahati
Book Image

OpenNI Cookbook

By: Soroush Falahati

Overview of this book

The release of Microsoft Kinect, then PrimeSense Sensor, and Asus Xtion opened new doors for developers to interact with users, re-design their application’s UI, and make them environment (context) aware. For this purpose, developers need a good framework which provides a complete application programming interface (API), and OpenNI is the first choice in this field. This book introduces the new version of OpenNI. "OpenNI Cookbook" will show you how to start developing a Natural Interaction UI for your applications or games with high level APIs and at the same time access RAW data from different sensors of different hardware supported by OpenNI using low level APIs. It also deals with expanding OpenNI by writing new modules and expanding applications using different OpenNI compatible middleware, including NITE. "OpenNI Cookbook" favors practical examples over plain theory, giving you a more hands-on experience to help you learn. OpenNI Cookbook starts with information about installing devices and retrieving RAW data from them, and then shows how to use this data in applications. You will learn how to access a device or how to read data from it and show them using OpenGL, or use middleware (especially NITE) to track and recognize users, hands, and guess the skeleton of a person in front of a device, all through examples.You also learn about more advanced aspects such as how to write a simple module or middleware for OpenNI itself. "OpenNI Cookbook" shows you how to start and experiment with both NIUI designs and OpenNI itself using examples.
Table of Contents (14 chapters)
OpenNI Cookbook
About the Author
About the Reviewers

Detecting a user's pose

In this recipe, we are going to show you how to request a search for a specific pose on a user and show the status of all the users' poses.

In the current version of NiTE, there are only two predefined poses that can be tracked and recognized: the PSI pose that was formally used as a calibration pose and the crossed hands pose that is a newly introduced pose. What follows is an image of a PSI pose:

In the new version of NiTE, there is no practical need to find out if a user is in one of these two predefined poses, because there is no need to be in a PSI pose for calibration and no requirement for the crossed hands pose that we are aware of. But you can still use these poses alone or if you want to support other third-party middleware. Also, even when it seems there is no need for PSI poses for calibration, it is still an error to indicate that calibration failed because of a lack of poses. So it may be used in some rare cases.

Getting ready

Create a project in Visual...