Complete visual metrology using relative affine structure
Summary (2 min read)
Introduction
- In computer vision, invariants are widely used for recognition and classification of objects and three dimensional reconstruction of a scene from one or more uncalibrated images [1][2][3][4].
- Broadly interpreted, all these vision tasks use invariants for retrieving geometric properties of objects from images.
- The camera model used in this work is the central projection.
- The relative affine structure is one of the widely used tools in the context of repeated objects [4].
- The method to compute measurements of repeated objects, individually, without using relative affine structures for corresponding points is developed in section III.
B. Apparatus for Proposed Framework
- The chosen world coordinate system for repeated objects is shown in figure (2).
- The three orthogonal planes πY Z , πZX and πXY are reference planes.
- Second constant (say µx0) is fixed for every point with respect to the reference plane (say πY Z).
- The proposed framework use this fundamental concept behind relative affine structure to determine 3D measurements of translaionally and affinely repeated objects.
- Any one cuboid is considered as principal object and rest as auxiliary objects.
III. MEASUREMENTS OF INDIVIDUAL OBJECT
- As described in section II-B, for a point Mi, a relative affine structure kx has a fixed constant µx0 and a variable constant 1 λi .
- The fixed constant will be eliminated by taking ratio of two relative affine structures for two different points with respect to same reference plane, πY Z .
- Thus, for any arbitrary point’s X coordinate can be computed using relative affine structure and projective depth.
- Similarly, the authors can compute the Y and Z coordinates of every point, given reference metric measurements Yref and Zref along Y and Z direction, respectively.
IV. PRINCIPAL OBJECT AS REFERENCE
- Alternately, the authors can arbitrarily choose a pair of points on two repeated objects.
- The respective relative affine structures kxi and k′xi for mi and m ′ i can be computed by Eq. (5) and Eq. (7).
- The ratio of kxi and k′xi can be expressed as follows kxi k′xi =.
V. TRANSFORMATION OF REPEATED OBJECT
- Under specific configurations, relative affine structure, which is projective structure, turns into affine structure.
- If the reference plane is at infinity or in case of parallel projection, relative affine structure approaches to affine structure [2].
- Thus, that ratio is a projective structure.
- Suppose S and S′ are repeated objects and are related by S′ = T (S), where T is a 4×4 general transformation matrix.
C. Affine Repetition
- The most general case is affine repetition that encapsulates rotation, translation, scaling and shearing [10].
- Once constants ψx0, ψy0 and ψz0 along X , Y and Z directions are computed from image (coordinates of M0), affine repetition can be computed uniquely.
VI. RESULTS
- In their experiments, the authors consider a real image, as shown in figure (3).
- Table I and II display the measurements of objects computed using methods discussed in sections III and IV, respectively.
- Based on the precision required for an application, the errors can be further reduced by employing efficient techniques for computing point correspondences and vanishing points from image.
- Additionally, proper uncertainly analysis will also improve the results [6].
VII. CONCLUSION
- The authors extended prior work on relative affine structure for computing three dimensional measurement from a single perspective image of repeated objects.
- The transformation between repeated objects can be represented in terms of relative affine structures along three orthogonal directions.
- Therefore, one invariant is used to analyze projective, affine and Euclidean space for vision tasks.
- Furthermore, three dimensional motion of an object or a camera can be parameterized in terms of relative affine structure.
- So, motion analysis related tasks such as motion segmentation and tracking can use relative affine structure, an invariant.
Did you find this useful? Give us your feedback
Citations
1 citations
Cites methods or result from "Complete visual metrology using rel..."
...Table I displays the measurements (in centimeters) computed using methods discussed in section II and III and compares the error percentage results with previous works [4][7]....
[...]
...In single view metrology, the uncertainty is added to the calculations and to metric values as the method considers vanishing point and vanishing line, hence the errors are inevitable because these parameters are assumed to be at infinity [4][5][7]....
[...]
References
14,282 citations
"Complete visual metrology using rel..." refers background or methods in this paper
...And, homology, a plane projective transformation, was used for vanishing points based visual metrology techniques and camera calibration [6][7][10][11][12]....
[...]
...This is a requirement to establish homology between two parallel planes [10]....
[...]
760 citations
"Complete visual metrology using rel..." refers background or methods in this paper
...And, homology, a plane projective transformation, was used for vanishing points based visual metrology techniques and camera calibration [6][7][10][11][12]....
[...]
...In this paper, we blend and develop previous results on relative affine structure and single view metrology [4][6][7][8][9]....
[...]
...In single view metrology, we can compute the distance between two parallel planes when the corresponding points on the planes are along the direction normal to the planes [7]....
[...]
249 citations
"Complete visual metrology using rel..." refers methods in this paper
...And, homology, a plane projective transformation, was used for vanishing points based visual metrology techniques and camera calibration [6][7][10][11][12]....
[...]
170 citations
"Complete visual metrology using rel..." refers methods in this paper
...And, homology, a plane projective transformation, was used for vanishing points based visual metrology techniques and camera calibration [6][7][10][11][12]....
[...]
167 citations
Related Papers (5)
Frequently Asked Questions (10)
Q2. What is the affine repetition of objects?
Once constants ψx0, ψy0 and ψz0 along X , Y and Z directions are computed from image (coordinates of M0), affine repetition can be computed uniquely.
Q3. What is the affine structure of a point?
Xi X ′i λ′i λi ψx0 ∼= Xi X ′i λ′i λi(14)Since ψx0 = λi0λ′ i0 X′i0 Xi0 is fixed for all points, the ratio kxik′ xi can be computed up to uniform scale along X-axis.
Q4. What is the relative affine structure of a point?
If the reference plane is at infinity or in case of parallel projection, relative affine structure approaches to affine structure [2].
Q5. What is the relative affine structure of a plane?
the relative affine structure is defined as [2],k = X1 λ1 λ0 X0(3)where λ1 and λ0 are depth of point M1 and M0 and X1 and X0 are perpendicular distance of these points from reference plane π which is Y Z plane in this case.
Q6. what is the affine structure of a point?
For every point M1 /∈ {πY Z , πZX , πXY } will have three relative affine structures, kx, ky and kz , respectively.kx = X1 1λ λ0 X0 , ky = Y1 1 λ λ0 X0 , and kz = Z1 1 λ λ0 X0(8)where, X0 and λ0 are X coordinate and depth of a fixed point M0 /∈ {πXY , πY Z , πZX}.
Q7. What is the relative affine structure of a camera?
The transformationbetween repeated objects can be represented in terms of relative affine structures along three orthogonal directions.
Q8. What is the ratio of relative affine structure to euclidean distance?
If object S undergoes pure rotation and results in S′, the transformation is represented as,X ′Y ′ Z ′ = r1 r2 r3r4 r5 r6 r7 r8 r9 XY Z = αx 0 00 αy 0 0 0 αz XY Z Therefore, the rotation between two points is equivalent to the following diagonal matrixR = αx 0 00 αy 0 0 0 αz where three columns are scaled uniformly along X , Y and Z directions, respectively.
Q9. What is the relation between S and M?
S′ is corresponding to M ∈ S. By using Eq. (16) and (17), the relation between corresponding points can be written as follows, X ′ Y ′Z ′1 = αx 0 0 0 0 αy 0 0 0 0 αz 0 0 0 0 1 X Y Z 1 = K X Y Z 1 (18) where K is the transformation matrix in 3D Euclidean space.
Q10. how can i compute a ratio ′ i?
The ratio λ ′ i λi can be computed by solving equation (4).λ′i λi = ||((Hπmi)× e′N )|| ||m′i ×mi||(15)Equation (14) can be written asX ′i ∼= Xi k′xi kxi λ′i λi = Xiαxi (16)Similarly, the authors can write expressions for Y and Z directions as below,Y ′i ∼= Yi k′yi kyi λ′i λi = Yiαyi, Z ′ i ∼= Zi k′zi kzi λ′i λi = Ziαzi (17)Given reference measurements on the principal object along X , Y and Z axes, measurements of affinely repeated object can be computed up to a respective scale by Eq. (16) and (17).