Abstract:We introduce several modifications of conic fitting in Geometric algebra for conics by incorporating additional conditions into the optimisation problem. Each of these extra conditions ensure additional geometric properties of a fitted conic, in particular, centre point position at the origin of coordinate system, axial alignment with coordinate axes, or, eventually, combination of both. All derived algorithms are accompanied by a discussion of the underlying algebra and optimisation issues, together with the implementation in MATLAB. Finally, we present examples on a sample dataset and offer possible use of the algorithms.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: in the framework of Geometric Algebra (GA), when fitting conic sections, how to optimize the fitting algorithm by introducing additional geometric constraints. Specifically, the author proposes three different geometric constraints to ensure that the fitted conic sections have specific geometric properties, such as the center point being located at the origin of the coordinate system, the axes being aligned with the coordinate axes, or satisfying both of these simultaneously.
### Specific description of the problem
1. **Background and motivation**:
- Conic section fitting is a common geometric problem and has wide applications in fields such as computer vision and image processing.
- Traditional conic section fitting methods are usually based on linear or quadratic constraints and can achieve optimal fitting by solving the eigenvectors of a matrix.
- Geometric Algebra (GA) provides a new method for representing and computing geometric objects and can handle geometric operations such as intersections, tangents, and distances more efficiently.
2. **Limitations of existing methods**:
- Although existing conic section fitting algorithms are effective, in some application scenarios, the fitted conic sections need to have specific geometric properties, such as the position of the center point or the alignment of the axes.
- Traditional methods are difficult to directly impose these geometric constraints, resulting in fitting results that may not meet the actual requirements.
3. **Objectives of the paper**:
- Introduce additional geometric constraints, such as the center point being located at the origin of the coordinate system and the axes being aligned with the coordinate axes, to improve the conic section fitting algorithm.
- Explicitly incorporate these constraints into the optimization problem in the framework of Geometric Algebra, thereby obtaining the optimal fitting results with specific geometric properties.
- Provide detailed mathematical derivations and MATLAB code implementations to verify the effectiveness of the proposed method.
### Specific geometric constraints
The three geometric constraints proposed in the paper are respectively:
1. **Axial Alignment**:
- The principal axis of the fitted conic section is aligned with the coordinate axes.
- Mathematically, this requires that the coefficient of the cross - term in the conic section equation is zero, that is, \( B = 0 \).
2. **Centre Point at Origin**:
- The center point of the fitted conic section is located at the origin of the coordinate system.
- Mathematically, this requires that the coefficients of the linear terms in the conic section equation are zero, that is, \( D = 0 \) and \( E = 0 \).
3. **Satisfying both Axial Alignment and Centre Point at Origin simultaneously**:
- Combine the above two conditions to ensure that the fitted conic section is both axially aligned and has its center point at the origin.
### Solution
The author solves the above problems through the following steps:
1. **Introducing geometric constraints**:
- On the basis of the original conic section fitting algorithm, introduce additional geometric constraints.
- These conditions are transformed into specific formulas and matrix expressions in Geometric Algebra.
2. **Redefining the optimization problem**:
- Redefine the conic section fitting problem with geometric constraints as an optimization problem.
- Use the Lagrange multiplier method to solve the constrained optimization problem and obtain the optimal solution.
3. **Algorithm implementation**:
- Provide detailed MATLAB code implementation to verify the effectiveness and practicality of the proposed method.
- Demonstrate the fitting results under different constraint conditions through experimental data and prove the feasibility and superiority of the method.
In summary, this paper improves the conic section fitting algorithm based on Geometric Algebra by introducing additional geometric constraints, so that the fitting results can better meet the specific geometric requirements in practical applications.