Before getting Kinect in motion, let's try to understand what motion computing (or motion control computing) is and how Kinect built its success in this area.
Motion control computing is the discipline that processes, digitalizes, and detects the position and/or velocity of people and objects in order to interact with software systems.
Motion control computing has been establishing itself as one of the most relevant techniques for designing and implementing a Natural User Interface (NUI).
NUIs are human-machine interfaces that enable the user to interact in a natural way with software systems. The goals of NUIs are to be natural and intuitive. NUIs are built on the following two main principles:
The NUI has to be imperceptible, thanks to its intuitive characteristics: (a sensor able to capture our gestures, a microphone able to capture our voice, and a touch screen able to capture our hands' movements). All these interfaces are imperceptible to us because their use is intuitive. The interface is not distracting us from the core functionalities of our software system.
The NUI is based on nature or natural elements. (the slide gesture, the touch, the body movements, the voice commands—all these actions are natural and not diverting from our normal behavior).
NUIs are becoming crucial for increasing and enhancing the user accessibility for software solution. Programming a NUI is very important nowadays and it will continue to evolve in the future.
Kinect embraces the NUIs principle and provides a powerful multimodal interface to the user. We can interact with complex software applications and/or video games simply by using our voice and our natural gestures. Kinect can detect our body position, velocity of our movements, and our voice commands. It can detect objects' position too.
Microsoft started to develop Kinect as a secret project in 2006 within the Xbox division as a competitive Wii killer. In 2008, Microsoft started Project Natal, named after the Microsoft General Manager of Incubation Alex Kipman's hometown in Brazil. The project's goal was to develop a device including depth recognition, motion tracking, facial recognition, and speech recognition based on the video recognition technology developed by PrimeSense.
Kinect for Xbox was launched in November 2010 and its launch was indeed a success: it was and it is still a break-through in the gaming world and it holds the Guinness World Record for being the "fastest selling consumer electronics device" ahead of the iPhone and the iPad.
In December 2010, PrimeSense (primesense.com) released a set of open source drivers and APIs for Kinect that enabled software developers to develop Windows applications using the Kinect sensor.
Finally, on June 17 2011 Microsoft launched the Kinect SDK beta, which is a set of libraries and APIs that enable us to design and develop software applications on Microsoft platforms using the Kinect sensor as a multimodal interface.
With the launch of the Kinect for Windows device and the Kinect SDK, motion control computing is now a discipline that we can shape in our garages, writing simple and powerful software applications ourselves.
This book is written for all of us who want to develop market-ready software applications using Kinect for Windows that can track audio and video and control motion based on NUI. In an area where Kinect established itself in such a short span of time, there is the need to consolidate all the technical resources and develop them in an appropriate way: this is our zero-to-hero Kinect in motion journey. This is what this book is about.
This book assumes that you have a basic knowledge of C# and that we all have a great passion to learn about programming for Kinect devices. This book can be enjoyed by anybody interested in knowing more about the device and learning how to track audio and video using the Kinect for Windows Software Development Kit (SDK) 1.6. We deeply believe this book will help you to master how to process video depth and audio stream and build market-ready applications that control motion. This book has deliberately been kept simple and concise, which will aid you to quickly grasp the core and critical concepts.
Before jumping on the core of audio and visual tracking with Kinect for Windows, let's take the space of this introduction chapter to understand what the hardware and software architectures Kinect for Windows and its SDK 1.6 use.
The Kinect device is a horizontal bar composed of multiple sensors connected to a base with a motorized pivot.
The following image provides a schematic representation of all the main Kinect hardware components. Looking at the Kinect sensor from the front, from the outside it is possible to identify the Infrared (IR) Projector (1), the RGB camera (3), and the depth camera (2). An array of four microphones (6), the three-axis accelerometer (5), and the tilt motor (4) are arranged inside the plastic case.
Kinect case and components
The device is connected to a PC through a USB 2.0 cable. It needs an external power supply in order to work because USB ports don't provide enough power.
Now let's jump in to the main features of its components.
The IR projector is the device that Kinect uses for projecting the IR rays that are used for computing the depth data. The IR projector, which from the outside looks like a common camera, is a laser emitter that constantly projects a pattern of structured IR dots at a wavelength around of 830 nm (patent US20100118123, Prime Sense Ltd.). This light beam is invisible to human eyes (that typically respond to wavelengths from about 390 nm to 750 nm) except for a red bright dot in the center of emitter.
The pattern is composed by 3 x 3 subpatterns of 211 x 165 dots (for a total of 633 x 495 dots). In each subpattern, one spot is much brighter than all the others.
As the dotted light (spot) hits an object, the pattern becomes distorted, and this distortion is analyzed by the depth camera in order to estimate the distance between the sensor and the object itself.
Infrared pattern
Note
In the previous image, we tested the IR projector against the room's wall. In this case we have to notice that a view of the clear infrared pattern can be obtained only by using an external IR camera (the left-hand side of the previous image). Taking the same picture from the internal RGB camera, the pattern will look distorted even though in this case the beam is not hitting any object (the right-hand side of the previous picture).
The depth camera is a (traditional) monochrome CMOS (complementary metal-oxide-semiconductor) camera that is fitted with an IR-pass filter (which is blocking the visible light). The depth camera is the device that Kinect uses for capturing the depth data.
The depth camera is the sensor returning the 3D coordinates (x, y, z) of the scene as a stream. The sensor captures the structured light emitted by the IR projector and the light reflected from the objects inside the scene. All this data is converted in to a stream of frames. Every single frame is processed by the PrimeSense chip that produces an output stream of frames. The output resolution is upto 640 x 480 pixels. Each pixel, based on 11 bits, can represent 2048 levels of depth.
The following table lists the distance ranges:
|
Mode |
Physical limits |
Practical limits |
|---|---|---|
|
Near |
0.4 to 3 m (1.3 to 9.8 ft) |
0.8 to 2.5 m (2.6 to 8.2 ft) |
|
Normal |
0.8 to 4 m (2.6 to 13.1 ft) |
1.2 to 3.5 m (4 to 11.5 ft) |
Note
The sensor doesn't work correctly within an environment affected by sunlight, a reflective surface, or an interference with light with a similar wavelength (830 nm circa).
The following figure is composed of two frames extracted from the depth image stream: the one on the left represents a scene without any interference. The one on the right is stressing how interference can reduce the quality of the scene. In this frame, we introduced an infrared source that is overlapping the Kinect's infrared pattern.
Depth images
The RGB camera is similar to a common color webcam, but unlike a common webcam, the RGB camera hasn't got an IR-cut filter. Therefore in the RGB camera, the IR is reaching the CMOS. The camera allows a resolution upto 1280 x 960 pixels with 12 images per second speed. We can reach a frame rate of 30 images per second at a resolution of 640 x 480 with 8 bits per channel producing a Bayer filter output with a RGGBD pattern. This camera is also able to perform color flicker avoidance, color saturation operations, and automatic white balancing. This data is utilized to obtain the details of people and objects inside the scene.
The following monochromatic figure shows the infrared frame captured by the RGB camera:
IR frame from the RGB camera
The Kinect cameras have a horizontal field of view of 57.5 degrees and a vertical field of view of 43.5 degrees. It is possible to increase the interaction space by adjusting the vertical tilt of the sensor by +27 and -27 degrees. The tilt motor can shift the Kinect head's angle upwards or downwards.
The Kinect also contains a three-axis accelerometer configured for a 2g range (g is the acceleration value due to gravity) with a 1 to 3 degree accuracy. It is possible to know the orientation of the device with respect to gravity reading the accelerometer data.
The following figure shows how the field of view angle can be changed when the motor is tilted:
Field of view angle
The microphone array consists of four microphones that are located in a linear pattern in the bottom part of the device with a 24-bit Analog to Digital Converter (ADC). The captured audio is encoded using Pulse Code Modulation (PCM) with a sampling rate of 16 KHz and a 16-bit depth. The main advantages of this multi-microphones configuration is an enhanced Noise Suppression, an Acoustic Echo Cancellation (AEC), and the capability to determine the location and the direction of an audio source through a beam-forming technique.
In this paragraph we review the software architecture defining the SDK. The SDK is a composite set of software libraries and tools that can help us to use the Kinect-based natural input. The Kinect senses and reacts to real-world events such as audio and visual tracking. The Kinect and its software libraries interact with our application via the NUI libraries, as detailed in the following figure:
Interaction diagram
Here, we define the software architecture diagram where we encompass the structural elements and the interfaces by which the Kinect for Windows SDK 1.6 is composed, as well as the behavior as specified in collaboration with those elements:
Kinect for Windows SDK 1.6 software architecture diagram
The following list provides the details for the information shown in the preceding figure:
Kinect sensor: The hardware components as detailed in the previous paragraph, and the USB hub through which the Kinect sensor is connected to the computer.
Kinect drivers: The Windows drivers for the Kinect, which are installed as part of the SDK setup process. The Kinect drivers are accessible in the
%Windows%\System32\DriverStore\FileRepositorydirectory and they include the following files:kinectaudio.inf_arch_uniqueGUID;kinectaudioarray.inf_arch_uniqueGUID;kinectcamera.inf_arch_uniqueGUID;kinectdevice.inf_arch_uniqueGUID;kinectsecurity.inf_arch_uniqueGUID
These files expose the information of every single Kinect's capabilities. The Kinect drivers support the following files:
The Kinect microphone array as a kernel-mode audio device that you can access through the standard audio APIs in Windows
Audio and video streaming controls for streaming audio and video (color, depth, and skeleton)
Device enumeration functions that enable an application to use more than one Kinect
Audio and video components defined by NUI APIs for skeleton tracking, audio, and color and depth imaging. You can review the NUI APIs header files in the
%ProgramFiles%\Microsoft SDKs\Kinect\v1.6folder as follows:NuiApi.h: This aggregates all the NUI API headersNuiImageCamera.h: This defines the APIs for the NUI image and camera servicesNuiSensor.h: This contains the definitions for the interfaces as theaudiobeam, theaudioarray, and the acceleratorNuiSkeleton.h: This defines the APIs for the NUI skeleton
DirectX Media Object (DMO) for microphone array beam-forming and audio source localization. The format of the data used in input and output by a stream in a DirectX DMO is defined by the
Microsoft.Kinect.DMO_MEDIA_TYPEand theMicrosoft.Kinect.DMO_OUTPUT_DATA_BUFFERstructs. The default facadeMicrosoft.Kinect.DmoAudioWrappercreates a DMO object using a registered COM server, and calls native DirectX DMO layer directly.Windows 7 standard APIs: The audio, speech, and media APIs in Windows 7, as described in the Windows 7 SDK and the Microsoft Speech SDK (
Microsoft.Speech,System.Media, and so on). These APIs are also available to desktop applications in Windows 8.
The stream of color image data is handled by the Microsoft.Kinect.ColorImageFrame. A single frame is then composed of color image data. This data is available in different resolutions and formats. You may use only one resolution and one format at a time.
The following table lists all the available resolutions and formats managed by the Microsoft.Kinect.ColorImageFormat struct:
|
Color image format |
Resolution |
FPS |
Data |
|---|---|---|---|
|
|
640 x 480 |
30 |
Pixel format is gray16 |
|
|
1280 x 960 |
12 |
Bayer data |
|
|
640 x 480 |
30 |
Bayer data |
|
|
640 x 480 |
15 |
Raw YUV |
|
|
1280 x 960 |
12 |
RGB (X8R8G8B8) |
|
|
640 x 480 |
15 |
Raw YUV |
|
|
N/A |
N/A |
N/A |
Note
When we use the InfraredResoluzion640x480Fps30 format in the byte array returned for each frame, two bytes make up one single pixel value. The bytes are in little-endian order, so for the first pixel, the first byte is the least significant byte (with the least significant 6 bits of this byte always set to zero), and the second byte is the most significant byte.
The X8R8G8B8 format is a 32-bit RGB pixel format, in which 8 bits are reserved for each color.
Raw YUV is a 16-bit pixel format. While using this format, we can notice the video data has a constant bit rate, because each frame is exactly the same size in bytes.
In case we need to increase the quality of the default conversion done by the SDK from Bayer to RGB, we can utilize the Bayer data provided by the Kinect and apply a customized conversion optimized for our central processing units (CPUs) or graphics processing units (GPUs).
Note
Due to the limited transfer rate of USB 2.0, in order to handle 30 FPS, the images captured by the sensor are compressed and converted in to RGB format. The conversion takes place before the image is processed by the Kinect runtime. This affects the quality of the images themselves.
In the SDK 1.6 we can customize the camera settings for optimizing and adapting the color camera for our environment (when we need to work in a low light or a brightly lit scenario, adapt contrast, and so on). To manage the code the Microsoft.Kinect.ColorCameraSettings class exposes all the settings we want to adjust and customize.
Note
In native code we have to use the Microsoft.Kinect.Interop.INuiColorCameraSettings interface instead.
In order to improve the external camera calibration we can use the IR stream to test the pattern observed from both the RGB and IR camera. This enables us to have a more accurate mapping of coordinates from one camera space to another.
The data provided by the depth stream is useful in motion control computing for tracking a person's motion as well as identifying background objects to ignore.
The depth stream is a stream of data where in each single frame the single pixel contains the distance (in millimeters) from the camera itself to the nearest object.
The depth data stream Microsoft.Kinect.DepthImageStream by the Microsoft.Kinect.DepthImageFrame exposes two distinct types of data:
Depth data calculated in millimeters (exposed by the
Microsoft.Kinect.DepthImagePixelstruct).Player segmentation data. This data is exposed by the
Microsoft.Kinect.DepthImagePixel.PlayerIndexproperty, identifying the unique player detected in the scene.
The following table defines the characteristics of the depth image frame:
|
Depth image format |
Resolution |
Frame rate |
|---|---|---|
|
|
640 x 480 |
30 FPS |
|
|
320 x 240 |
30 FPS |
|
|
80 x 60 |
30 FPS |
|
|
N/A |
N/A |
The Kinect runtime processes depth data to identify up to six human figures in a segmentation map. The segmentation map is a bitmap of Microsoft.Kinect.DepthImagePixel, where the
PlayerIndex property identifies the closest person to the camera in the field-of-view. In order to obtain player segmentation data, we need to enable the skeletal stream tracking.
Microsoft.Kinect.DepthImagePixel has been introduced in the SDK 1.6 and defines what is called the "Extended Depth Data", or
full depth information: each single pixel is represented by a 16-bit depth and a 16-bit player index.
Thanks to the microphone array, the Kinect provides an audio stream that we can control and manage in our application for audio tracking, voice recognition, high-quality audio capturing, and other interesting scenarios.
By default, Kinect tracks the loudest audio input. Having said that, we can certainly direct programmatically the microphone arrays (towards a given location, or following a tracked skeleton, and so on).
DirectX Media Object (DMO) is the building block used by Kinect for processing audio streams.
Note
In native scenario in addition to the DirectX Media Object (DMO), we can use the Windows Audio Session API (WASAPI) too.
In managed applications, the Microsoft.Kinect.KinectAudioSource class (exposed in the KinectSensor.AudioSource property) is the key software architecture component concerning the audio stream. Using the Microsoft.Kinect.INativeAudioWrapper class wraps the DirectX Media Object (DMO), which is a common Windows component for a single-channel microphone.
The KinectAudioSource class is not limited to wrap the DMO, but it introduces additional abilities such as:
Audio input range
Note
In order to increase the quality of the sound, audio inputs coming from the sensor get upto a 20 dB suppression. The array microphone allows an optional additional 6 dB of ambient noise removal for audio coming from behind the sensor.
The audio input has a range of +/– 50 degrees (as visualized in preceding figure) in front of the sensor. We can point the audio direction programmatically using a 10 degree increment range in order to focus our attention on a given user or to elude noise sources.
In addition to the data provided by the depth stream, we can use those provided by the skeleton tracking to enhance the motion control computing capabilities of our applications in regards to recognizing people and following their actions.
We define the skeleton as a set of positioned key points. A detailed skeleton contains 20 points in normal mode and 10 points in seated mode, as shown in the following figure. Every single point of the skeleton highlights a joint of the human body.
Thanks to the depth (IR) camera, Kinect can recognize up to six people in the field of view. Of these, up to two can be tracked in detail.
The stream of skeleton data is maintained by the Microsoft.Kinect.SkeletonStream class and the Microsoft.Kinect.SkeletonFrame class. The skeleton data is exposed for each single point in the 3D space by the Microsoft.Kinect.SkeletonPoint struct. In any single frame handled by the skeleton stream we can manage up to six skeletons using an array of the Microsoft.Kinect.Skeleton class.
Skeleton in normal and seated mode
In this chapter we introduced Kinect, looking at the key architectural aspects such as the hardware composition and the SDK 1.6 software components. We walked through the color sensor, IR depth sensors, IR emitter, microphone arrays, the tilt motor for changing the Kinect camera angles, and the three-axis accelerometer.
Kinect generates two video streams using the color camera data and the depth information using the depth sensor. Kinect can detect up to six users in its view field and produce a detailed skeleton for two of them. All these characteristics make Kinect an awesome tool for video tracking motion. The Kinect's audio tracking makes the device a remarkable interface for voice recognition. Combining video and audio, Kinect and its SDK 1.6 are an outstanding technology for NUI.
Kinect is not just technology, it is indeed a means of how we can elevate the way users interact with complex software applications and systems. It is a break-through on how we can include NUIs and multimodal interface.
Kinect discloses unlimited opportunities to developers and software architects to design and create modern applications for different industries and lines of business.
The following examples are not meant to be an exhaustive list, but just a starting point that can inspire your creativity and increase your appetite for this technology.
Healthcare: This improves the physical rehabilitation process by constantly capturing data of the motion and posture of patient. We can enhance this scenario by allowing doctors to check the patient data remotely streamed by the Kinect sensor.
Education/Professional development: This helps in creating safe and more engaging environments based on gamification where students, teachers, and professionals can exercise activities and knowledge. The level of engagement can be increased even further using augmented reality.
Retail: This engages customers across multiple channels using the Kinect's multimodal interface. Kinect can be used as a navigation system for virtual windows while shopping online and/or visiting infotainment kiosks.
Home automation: This is also known as domotics where, thanks to the Kinect audio and video tracking, we can interact with all the electrical devices installed at our home (lights, washing machine, and so on).
In the next chapter, we will start to develop with the Kinect SDK, utilizing the depth and RGB camera streams. The applied examples will enable our application to optimize the way we manage and tune the streams themselves.
