CN105513128A - Kinect-based three-dimensional data fusion processing method - Google Patents

Abstract

The invention discloses a Kinect-based three-dimensional data fusion processing method, comprising the following steps: a, using two sets of Kinect to respectively acquire point cloud data A and B of the same chessboard target, and obtaining spatial measurement coordinates of 40 identical chessboard lattice points in the two groups of point clouds; symmetrically arranging the two sets of Kinect relative to the chessboard target; b, acquiring a conversion matrix Mwc under a space coordinates system in which a point cloud B converted from a point cloud A is located; c, performing data fusion on three-dimensional point clouds acquired by the two sets of Kinect. According to the Kinect-based three-dimensional data fusion processing method, two sets of Kinect equipment are used to acquire a group of three-dimensional point cloud data, and space positions of the two sets of Kinect are demarcated, thus the conversion matrix is obtained, and further the data fusion on the acquired two groups of three-dimensional point clouds is realized.

Description

Based on the three-dimensional data method for amalgamation processing of Kinect

Technical field

The present invention relates to the three-dimensional data method for amalgamation processing based on Kinect.

Background technology

In the optical measurement of wind tunnel model attitude and distortion, to paste, spray patterns mark or embed luminescent marking and destroy model surface characteristic, and be difficult to retain when temperature, pressure change is violent.In order to process especially model not needing when model attitude and displacement measurement in wind tunnel test, as spraying or additional marking point, raising, needs to study a kind of three-dimensional non-contact measurement method newly as the power of test under high-temperature and high-pressure conditions to harsh test environment.

Microsoft's Kinect device is a kind of degree of depth video camera, and it has imported the functions such as instant motion capture, image identification, microphone input, speech recognition, community interactive simultaneously.Do not need to use any controller, it is the 3 d pose and the deformation information that rely on the motion of model in cameras capture three dimensions to obtain tested model.

Although depth information collecting device acquisition precision popular is at present high, the condition required often compares Xun and carves, and adds the reason such as price and complicated operation, cannot reach civilian effect.Because the infrared pick-up head of Kinect and VGA camera are in different positions, and the parameter of camera lens itself is also incomplete same, so the picture acquired by two video cameras has slightly little difference, three-dimensional coordinate (X cannot be made, Y, Z) and the same point of the corresponding model of chromatic information.

Summary of the invention

The object of the present invention is to provide a kind of three-dimensional data method for amalgamation processing based on Kinect, can under complex environment, two Kinect device are utilized to obtain one group of three dimensional point cloud, and by demarcating two Kinect locus, obtain transition matrix, and then the two groups of three-dimensional point clouds obtained are realized the fusion of data.

For achieving the above object, technical scheme of the present invention is a kind of three-dimensional data method for amalgamation processing based on Kinect of design, comprises the steps:

A. obtain cloud data A and B of same chessboard target with two Kinect respectively, and obtain the space measurement coordinate of 40 identical checker-wise o'clock in two groups of some clouds; And two Kinect place with chessboard target symmetry;

B. acquisition point cloud A is transformed into the transition matrix M under a space coordinates at cloud B place ^wc, use least square method, according to formula M ^wc=(A ^ta) ^-1a ^tb obtains corresponding conversion matrix;

C. the three-dimensional point cloud that two Kinect obtain is carried out data fusion: be transformed into below the world coordinate system at another Kinect place by the cloud data that the Kinect obtaining A point cloud obtains, formulae express is as follows: AM ^wc=B.

Preferably, two Kinect place with chessboard target symmetry.

Advantage of the present invention and beneficial effect are: provide a kind of three-dimensional data method for amalgamation processing based on Kinect, can under complex environment, two Kinect device are utilized to obtain one group of three dimensional point cloud, and by demarcating two Kinect locus, obtain transition matrix, and then the two groups of three-dimensional point clouds obtained are realized the fusion of data.

The present invention uses the software write based on OpenNI and Primesense to obtain three-dimensional data points cloud in conjunction with Kinect, and the depth data (Z) that Kinect can be made to obtain and view data (X, Y) can be good at coincidence.

The present invention obtains model aximal deformation value, the 3 d pose in large interpretation region and deformation data when not damage model character of surface, the measuring table three-dimensional coordinate result of building and the consistance of portable three-coordinate instrument system better.The method can not only make its depth data and view data can be good at coincidence, but also the errorless combination of the cloud data that two Kinect can be obtained is fused together.

Accompanying drawing explanation

Fig. 1 is schematic diagram of the present invention;

Fig. 2 is target schematic diagram;

Fig. 3 is target calibration maps.

Embodiment

Below in conjunction with drawings and Examples, the specific embodiment of the present invention is further described.Following examples only for technical scheme of the present invention is clearly described, and can not limit the scope of the invention with this.

As shown in Figure 1 to Figure 3, the technical scheme that the present invention specifically implements is:

(1) computing machine (4 in Fig. 1) is used, the process software that built-in computer is write based on OpenNI and Primesense, two Kinect (1 in Fig. 1,2) are positioned over the position shown in Fig. 1, obtain target 3 front (as shown in Figure 2) tessellated 3 d space coordinate point respectively.

(2) two Kinect (1 in Fig. 1, the 2) spatial relation that front and back are placed is demarcated, and the transition matrix obtained; Two Kinect place with chessboard target symmetry; Key step comprises:

A. two Kinect (1 in Fig. 1,2) are used to obtain cloud data A and B of same chessboard target (3 in Fig. 1) respectively, and obtain identical 40 checker-wise points (as shown in Figure 3,5 in Fig. 3 is calibration point) the space measurement coordinate in two groups of some clouds with

B. the transformational relation that is transformed under a space coordinates at cloud B place of acquisition point cloud A is as follows:

$X_A^W=\frac{m_{11}^{wc}{X}_B^W+m_{12}^{wc}{Y}_B^W+m_{13}^{wc}{Z}_B^W+m_{14}^{wc}}{m_{41}^{wc}{X}_B^W+m_{42}^{wc}{Y}_B^W+m_{43}^{wc}{Z}_B^W+m_{44}^{wc}}---\left(1\right)$

$Y_A^W=\frac{m_{21}^{wc}{X}_B^W+m_{22}^{wc}{Y}_B^W+m_{23}^{wc}{Z}_B^W+m_{24}^{wc}}{m_{41}^{wc}{X}_B^W+m_{42}^{wc}{Y}_B^W+m_{43}^{wc}{Z}_B^W+m_{44}^{wc}}---\left(2\right)$

$Z_A^W=\frac{m_{31}^{wc}{X}_B^W+m_{32}^{wc}{Y}_B^W+m_{33}^{wc}{Z}_B^W+m_{34}^{wc}}{m_{41}^{wc}{X}_B^W+m_{42}^{wc}{Y}_B^W+m_{43}^{wc}{Z}_B^W+m_{44}^{wc}}---\left(3\right)$

Transition matrix is M ^wc, use least square method, according to formula M ^wc=(A ^ta) ^-1a ^tb obtains corresponding conversion matrix.M ^wcbe expressed as follows:

$M^{wc}=\left[\begin{array}{cccc}{m}_{11}^{wc}& m_{12}^{wc}& m_{13}^{wc}& m_{14}^{wc}\\ m_{21}^{wc}& m_{22}^{wc}& m_{23}^{wc}& m_{24}^{wc}\\ m_{31}^{wc}& m_{32}^{wc}& m_{33}^{wc}& m_{34}^{wc}\\ m_{41}^{wc}& m_{42}^{wc}& m_{43}^{wc}& m_{44}^{wc}\end{array}\right]---\left(4\right)$

(3) three-dimensional point cloud that two Kinect obtain is carried out data fusion, key step comprises, and be transformed into below the world coordinate system at another Kinect place by the cloud data that the Kinect obtaining A point cloud obtains, formulae express is as follows: $\left[\begin{array}{c}{X}_{B}w\\ Y_{B}w\\ Z_{B}w\\ w\end{array}\right]=M^{wc}\left[\begin{array}{c}{X}_A\\ Y_A\\ Z_A\\ 1\end{array}\right],$ Realize the comprehensive three-dimensional non-contact measurement to model.

The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the prerequisite not departing from the technology of the present invention principle; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (2)

1., based on the three-dimensional data method for amalgamation processing of Kinect, it is characterized in that, comprise the steps:

A. obtain cloud data A and B of same chessboard target with two Kinect respectively, and obtain the space measurement coordinate of 40 identical checker-wise o'clock in two groups of some clouds;

B. acquisition point cloud A is transformed into the transition matrix M under a space coordinates at cloud B place ^wc, use least square method, according to formula M ^wc=(A ^ta) ^-1a ^tb obtains corresponding conversion matrix;

C. the three-dimensional point cloud that two Kinect obtain is carried out data fusion: be transformed into below the world coordinate system at another Kinect place by the cloud data that the Kinect obtaining A point cloud obtains, formulae express is as follows: AM ^wc=B.

2. the three-dimensional data method for amalgamation processing based on Kinect according to claim 1, is characterized in that, two Kinect place with chessboard target symmetry.