Geodesic intersections

Charles F. F. Karney

DOI: https://doi.org/10.1061/JSUED2.SUENG-1483

2023-12-20

Abstract:A complete treatment of the intersections of two geodesics on the surface of an ellipsoid of revolution is given. With a suitable metric for the distances between intersections, bounds are placed on their spacing. This leads to fast and reliable algorithms for finding the closest intersection, determining whether and where two geodesic segments intersect, finding the next closest intersection to a given intersection, and listing all nearby intersections. The cases where the two geodesics overlap are also treated.

Geophysics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **How to find the intersection points of two geodesics on a rotating ellipsoid, especially the nearest intersection point, and provide a fast and reliable method to deal with the related problems of intersection points**.

### Problem Background
1. **Geodesic Intersection Point Problem**:
- A geodesic is the shortest path between two points on an ellipsoid. On a sphere, a geodesic corresponds to a great circle.
- On a rotating ellipsoid, geodesics are usually not closed curves, so they will intersect repeatedly at different points.
- The goal of the paper is to systematically study all the intersection points of two geodesics on a rotating ellipsoid and pay special attention to how to find the nearest intersection point.

2. **Deficiencies of Existing Methods**:
- The method of Sjöberg (2008) is too complicated because it combines the geodesic problem and the intersection point problem.
- Although the method of Baselga and Martínez - Llario (BML, 2018) provides a solution, it has the following problems:
- The solution to the spherical intersection point problem is wrong.
- It cannot guarantee to find the nearest intersection point.

3. **Goals of the Paper**:
- Correct the errors in the BML method.
- Provide a reliable method to find the nearest intersection point.
- Deal with other related problems, such as:
- Judging whether two geodesic segments intersect and the position of the intersection.
- Finding the next nearest intersection point after a given intersection point.
- Listing all the intersection points within a certain range.

---

### Specific Problem Decomposition

1. **Define "Nearest Intersection Point"**:
- Use the \( L_1 \) distance metric to define the "nearness" of the intersection point:
\[
L_1(S)=|x| + |y|
\]
- Here, \( S = [x, y] \) represents the displacement vectors on the two geodesics.

2. **Distance Limitations of Intersection Points**:
- Find the minimum distance \( t_1 \) and the maximum distance \( t_2 \) between intersection points.
- Ensure that the basic algorithm can find the nearest intersection point within a capture radius \( t_4 \).

3. **Algorithm Design**:
- **Nearest Intersection Point Algorithm**:
- Use the basic algorithm to start the search from multiple starting points.
- Screen out the nearest intersection point according to the \( L_1 \) distance.
- **Geodesic Segment Intersection Point Problem**:
- Judge whether the intersection point falls within the rectangular range of the two geodesic segments.
- **Other Intersection Point Problems**:
- Find the next nearest intersection point after a given intersection point.
- List all the intersection points within a certain range.

4. **Handling Special Cases**:
- **Coincident Geodesics**:
- When two geodesics are almost coincident, multiple intersection points may occur.
- These cases need to be specially handled to ensure the correctness and efficiency of the algorithm.

---

### Summary

The core problem of the paper is: **How to efficiently and reliably find the nearest intersection point of two geodesics on a rotating ellipsoid and extend it to other related problems**. By introducing the \( L_1 \) distance metric and a series of optimization algorithms, the author has solved the deficiencies of the existing methods and provided a complete solution to the geodesic intersection point problem.