"Translation" a thorough comparison of Kinect v1 and Kinect v2

"Original: Http://www.cnblogs.com/TracePlus/p/4136297.html"

This series is primarily a comparison of the current version of Kinect for Windows (v1) and the Next Generation Developer Preview (v2) to introduce the evolving hardware and software in the context of C + + developers.           This paper is a thorough comparison of the configuration and operating conditions of the sensor.   The Kinect for Windows Developer Preview presented in this series is tentative, and the software, hardware, and APIs may be subject to change due to the final product version, please understand. About this serial series is mainly the comparison of the next generation of Kinect for Windows (later called Kinect v2 Preview) and the current type of Kinect for Windows (later called Kinect v1), while introducing the evolution of the C + + developers Hardware and software. (This site also publishes content that corresponds to the C#/visual basic developer.) NET developers can reference [new Kinect for Windows v2 Developer Preview programing]
The next generation of Kinect for Windows 2012 US Microsoft released the Kinect v1, because it is easy to get depth (depth) and skeleton (character posture) and other information, by developers and researchers around the world.     The Kinect v2, which was scheduled for sale in 2014, predicts a great evolution in hardware and software that was released early in the preview version of the Kinect v2 (sensor) and SDK V2 (software Development Kit) before the sale.   Also, because this developer to the early provider is required to sign an NDA (password retention contract), this article is not published in advance please understand the matter.     The appearance of the Kinect v1 and Kinect v2 Preview compare to the appearance of the Kinect v1 (Figure 1) and Kinect v2 Preview (Figure 2). Figure 1 Kinect for Windows v1 (current type) Kinect v 1 depth sensor, using the "light coding" method, read the projected infrared pattern, through pattern deformation to obtain depth information. For this reason, the depth sensor is divided into IR projector (left) that casts infrared pattern and the IR Camera (right) to read this. There is also a color Camera in the middle of the depth sensor.
Light coding is a depth sensor technology from Israel's PrimeSense company, which was acquired by Apple in the United States in 2013. Figure 2 Kinect for Windows v2 (Next generation) preview of the Kinect V2 Preview version of the depth sensor, using "time of Flight (TOF), to obtain depth information by returning time from the projected infrared reflection.        The depth sensor does not see the appearance, but the color camera is next to the infrared camera (left) and the Porjector (right) of the projected pulse pitch infrared.   Microsoft has in the past acquired a company that uses TOF to process depth sensor technology (note: It should be referred to as 3DV) and is already using the technology, but without detailed disclosure. The configuration comparison of the Kinect v1 and Kinect v2 preview versions of the Kinect v1 and Kinect v2 preview is shown in table 1.

		Kinect v1	Kinect v2 Preview
colors (color)	Resolution (Resolution)	640x480	1920x1080
	Fps	30fps	30fps
Depth (Depth)	Resolution (Resolution)	The	512x424
	Fps	30fps	30fps
Number of people (Player)		6 people	6 people
Character Posture (Skeleton)		2 people	6 people
Shimonoseki (Joint)		20 Shimonoseki/person	25 Shimonoseki/person
The opening of the hand (Hand State)		(Developer Toolkit)	0 (SDK)
Detection Girth (Range of Detection)		0.8~4.0m	0.5~4.5m
Angle (Angle) (Depth)	level (horizontal)	57 degrees	70 degrees
	Vertical (Vertical)	43 degrees	60 degrees
(Tilt Motor)		0	X (Manual)
The number of apps		X (single one app)	0

  Table 1 is the configuration comparison of the Kinect v1 and Kinect v2 preview sensors   The Kinect v1 color camera has a low resolution of 640x480, does not get a very nice image, and the Kinect v2 preview has a significantly higher resolution To get a very nice image of 1920x1080 (Figure 3). (Note: The requirement of V1 is that the USB2.0 theoretical transfer rate is 60mb/s,v2 is USB3.0 the theoretical transfer rate is 500mb/s. You can calculate, in the case of X rgb  Color, 30fps, then the data required to be transferred per second is 640 x 480 x 4 x 30 is about 35M, plus the ushort format depth Color, 30fps, size is 2 x x x 30 is approximately 4M. The total is about 40mb/s, because bandwidth is limited, so in the case of guaranteed frame rate stability, the resolution can only be so, and basically must be exclusive of a USB Controller. In the case of v2, Color = 1920x1080 x 4 x 30 is about 237M, Depth = x 424 x 2 x 30 is about 12M, the total is about 250m/s. So it is not USB3.0, otherwise it cannot transmit such a large amount of data. Obviously, Color map is the most bandwidth, in fact, can be used in some other format, such as I420 or mjpg to reduce the amount of data, and then through the CPU or GPU for decompression and playback. )     Figure 3 Kinect v1 and Kinect v2 preview version of color