CN106918302A - A kind of spatial digitizer self-adapting calibration method - Google Patents

Abstract

The invention provides a kind of spatial digitizer self-adapting calibration method, it is characterised in that comprise the following steps：The point that Harris angle points are concentrated is divided three classes；It is every point of Harris angle points concentration, sets up its Laplce's coordinate；If 3 D scanning system there are n platform cameras, take on the basis of the Harris angle point collection that the 1st camera is gathered, the Harris angle points collection of other each camera carries out Automatic adjusument.Using after the method that the present invention is provided, the angle point overlap problem in three-dimensional scanner system marked point process can be solved, realize the self-adapting calibration to three-dimensional scanner system.

Description

A kind of spatial digitizer self-adapting calibration method

Technical field

The present invention relates to a kind of method that self-adapting calibration is carried out to three-dimensional scanner system.

Background technology

When three-dimensional scanner system is built using multiple depth cameras, it is necessary to the three-dimensional point cloud that single camera is gathered Piece is stitched together.Its splicing principle is that, using same scaling board, there is 16 × 10 chequered with black and white square dices above, The angle point for being constituted, referred to as Harris angle points.According to the image-forming principle of camera, if the purpose of camera calibration is by cadre's phase Several Harris angle point figures that machine is seen overlap exactly.However, because the processing technology of camera is not quite similar, And when three-dimensional scanner system is built, (such as camera case is under different screw tension force for the objective reality of camera rigging error Different degreeof tortuositys), angle point excessive effects can be caused.

The content of the invention

The purpose of the present invention is：Overcome angle point excessive effects, realize the self-adapting calibration to three-dimensional scanner system.

In order to achieve the above object, the technical scheme is that providing a kind of spatial digitizer self-adapting calibration side Method, it is characterised in that comprise the following steps：

Step 1, by Harris angle points concentrate point be divided three classes, I types point be Harris angle points concentrate non-boundary point And non-four angle points, Type-II point is the boundary point of non-four angle points that Harris angle points are concentrated, and type III point is Harris angles Four angle points that point is concentrated；

Step 2, it is point that every Harris angle point is concentrated, sets up its Laplce's coordinate, comprises the following steps：

Step 2.1, calculating projection plane P=n_xx+n_yy+n_zZ+d, its normal direction n are Wherein：

v_iIt is the nearest-neighbors point of point v, for I types point, possesses four nearest-neighbors points, k=4；For Type-II point, Possess three nearest-neighbors points, k=3；For type III point, possess two nearest-neighbors points, k=2；

D is average distance,

n_x、n_y、n_zIt is components of the normal direction n on tri- directions of xyz；

Step 2.2, by point v and its nearest-neighbors spot projection to projection plane P, the locus expression of its subpoint is such as Under：

v_proj=v- (d+ (vn)) n；

v_{i_proj}=v_i-(d+(v_iN)) n, in formula, v_projIt is projections of the point v on projection plane P, v_{i_proj}For its is nearest Projection of neighbours' point on projection plane P；

Step 2.3, calculating point v points are relative to v_iThe tangential weight w of point_iAnd normal direction weight b_i；

If step 3,3 D scanning system have n platform cameras, on the basis of taking the Harris angle point collection that the 1st camera is gathered, The Automatic adjusument process of the Harris angle point collection of other each camera is：

Step 3.1, for jth platform camera, j ≠ 1, Harris angle point collection for, according to the method in step 2.1, meter Calculate its projection plane being currently located, and average distance d ' and normal direction n '；

Step 3.2, for jth platform camera, j ≠ 1, Harris angle point collection in i-th point for, by its neighbours' point v′_iProject on new projection plane, gained projected position v '_{i_proj}For：

v′_{i_proj}=v '_i-[d′+(v′_i·n′)]n′

Step 3.3, the tangential weight w according to step 2.3_i, calculate jth platform camera, j ≠ 1, Harris angle point collection in I-th point of projected position v '_proj：

Step 3.4, calculate jth platform camera, j ≠ 1, Harris angle point collection in i-th point of final position

Preferably, tangential weight w_iCalculating process be：

In formula,Wherein：

Preferably, normal direction weight b_iCalculating process be：

In formula,

Asked using the overlap of the angle point in three-dimensional scanner system marked point process after the method that the present invention is provided, can be solved Topic, realizes the self-adapting calibration to three-dimensional scanner system.

Brief description of the drawings

Fig. 1 is using the effect contrast figure after the present invention.

Specific embodiment

With reference to specific embodiment, the present invention is expanded on further.It should be understood that these embodiments are merely to illustrate the present invention Rather than limitation the scope of the present invention.In addition, it is to be understood that after the content for having read instruction of the present invention, people in the art Member can make various changes or modifications to the present invention, and these equivalent form of values equally fall within the application appended claims and limited Scope.

The invention provides a kind of spatial digitizer self-adapting calibration method, comprise the following steps：

Step 1, by Harris angle points concentrate point be divided three classes, I types point be Harris angle points concentrate non-boundary point And non-four angle points, Type-II point is the boundary point of non-four angle points that Harris angle points are concentrated, and type III point is Harris angles Four angle points that point is concentrated；

Step 2, it is point that every Harris angle point is concentrated, sets up its Laplce's coordinate, comprises the following steps：

Step 2.1, calculating projection plane P=n_xx+n_yy+n_zZ+d, its normal direction n are Wherein：

v_iIt is the nearest-neighbors point of point v, for I types point, possesses four nearest-neighbors points, k=4；For Type-II point, Possess three nearest-neighbors points, k=3；For type III point, possess two nearest-neighbors points, k=2；

D is average distance,

n_x、n_y、n_zIt is components of the normal direction n on tri- directions of xyz；

Step 2.2, by point v and its nearest-neighbors spot projection to projection plane P, the locus expression of its subpoint is such as Under：

v_proj=v- (d+ (vn)) n；

v_{i_proj}=v_i-(d+(v_iN)) n, in formula, v_projIt is projections of the point v on projection plane P, v_{i_proj}For its is nearest Projection of neighbours' point on projection plane P；

Step 2.3, calculating point v points are relative to v_iThe tangential weight w of point_iAnd normal direction weight b_i；

Tangential weight w_iCalculating process be：

In formula,Wherein：

Normal direction weight b_iCalculating process be：

In formula,

If step 3,3 D scanning system have n platform cameras, on the basis of taking the Harris angle point collection that the 1st camera is gathered, The Automatic adjusument process of the Harris angle point collection of other each camera is：

Step 3.1, for jth platform camera, j ≠ 1, Harris angle point collection for, according to the method in step 2.1, meter Calculate its projection plane being currently located, and average distance d ' and normal direction n '；

Step 3.2, for jth platform camera, j ≠ 1, Harris angle point collection in i-th point for, by its neighbours' point v′_iProject on new projection plane, gained projected position v '_{i_proj}For：

v′_{i_proj}=v '_i-[d′+(v′_i·n′)]n′

Step 3.3, the tangential weight w according to step 2.3_i, calculate jth platform camera, j ≠ 1, Harris angle point collection in I-th point of projected position v '_proj：

Step 3.4, calculate jth platform camera, j ≠ 1, Harris angle point collection in i-th point of final position

Fig. 1 is shown and is compared using the effect of camera collection three-dimensional point cloud before and after the technology of the present invention, it can be seen that by this After inventive technique is demarcated to camera, high-quality three-dimensional point cloud can be generated.

Claims (3)

1. a kind of spatial digitizer self-adapting calibration method, it is characterised in that comprise the following steps：

Step 1, the point that Harris angle points are concentrated is divided three classes, I types point is the non-boundary point and non-that Harris angle points are concentrated Four angle points, Type-II point is the boundary point of non-four angle points that Harris angle points are concentrated, and type III point is Harris angle point collection In four angle points；

Step 2, it is point that every Harris angle point is concentrated, sets up its Laplce's coordinate, comprises the following steps：

Step 2.1, calculating projection plane P=n_xx+n_yy+n_zZ+d, its normal direction n are Wherein：

v_iIt is the nearest-neighbors point of point v, for I types point, possesses four nearest-neighbors points, k=4；For Type-II point, possess Three nearest-neighbors points, k=3；For type III point, possess two nearest-neighbors points, k=2；

D is average distance,

n_x、n_y、n_zIt is components of the normal direction n on tri- directions of xyz；

Step 2.2, by point v and its nearest-neighbors spot projection to projection plane P, the locus of its subpoint is expressed as follows：

v_proj=v- (d+ (vn)) n；

v_{i_proj}=v_i-(d+(v_iN)) n, in formula, v_projIt is projections of the point v on projection plane P, v_{i_proj}It is its nearest-neighbors Projection of the point on projection plane P；

Step 2.3, calculating point v points are relative to v_iThe tangential weight w of point_iAnd normal direction weight b_i；

If step 3,3 D scanning system have n platform cameras, on the basis of taking the Harris angle point collection that the 1st camera is gathered, other The Automatic adjusument process of each Harris angle point collection of camera is：

Step 3.1, for jth platform camera, j ≠ 1, Harris angle point collection for, according to the method in step 2.1, calculate it The projection plane being currently located, and average distance d ' and normal direction n '；

Step 3.2, for jth platform camera, j ≠ 1, Harris angle point collection in i-th point for, by its neighbours' point v '_iProjection Onto new projection plane, gained projected position v '_{i_proj}For：

v′_{i_proj}=v '_i-[d′+(v′_i·n′)]n′

Step 3.3, the tangential weight w according to step 2.3_i, calculate jth platform camera, j ≠ 1, Harris angle point collection in i-th The projected position v ' of point_proj：

$v_{proj}^{\′}=\underset{i=1}{\overset{k}{\Σ}}{w}_i{v}_{i\_proj}^{\′}$

Step 3.4, calculate jth platform camera, j ≠ 1, Harris angle point collection in i-th point of final position v ',

2. a kind of spatial digitizer self-adapting calibration method as claimed in claim 1, it is characterised in that in the step 1.3 In, tangential weight w_iCalculating process be：

In formula,Wherein：

3. a kind of spatial digitizer self-adapting calibration method as claimed in claim 1, it is characterised in that in the step 1.3 In, normal direction weight b_iCalculating process be：

In formula,