The Ptolemy-Alhazen problem with source at infinity

Masayo Fujimura, Matti Vuorinen · 2026

Read The Ptolemy-Alhazen problem with source at infinity by Masayo Fujimura, Matti Vuorinen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-25160.txt.

Introduction

Fujimura and Vuorinen revisit a classic optics puzzle traced to Ptolemy and Alhazen. The problem asks where a ray must strike a spherical mirror so that light from a source reflects to a given eye. Their case places the source at infinity, so incoming rays arrive parallel, like sunlight. This far-source version changes the geometry of the bounce point search. The pair studies the reflection problem in that limit. The setup links an ancient question to clean modern form.

Catalog record

CREATORMasayo Fujimura, Matti Vuorinen
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE18152
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.25160

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The Ptolemy-Alhazen problem with source at infinity Masayo Fujimura, Matti Vuorinen Abstract . We study the well-known Ptolemy-Alhazen problem on reflection of light at the surface of a spherical mirror in the case when the source of light is very far from the mirror. 2010 Mathematics Subject Classification 51M09, 51M15 † † footnotetext: File: paInfty20260326.tex, printed: 2026-8-10, 5.54 1. Introduction The well-known Ptolemy–Alhazen problem reads [ S ] , [ GBL , p.1010] : "Given a light source and a spherical mirror, find the point on the mirror where the light will be reflected to the eye of an observer." The long history of this problem is described in [ S , FMV2 , W ] . The problem is of current interest for instance in the design of computer vision of drones, in the study of GPS signal transmission, in many scattering problems of electromagnetic signals (e.g. radar signals), acoustics, and in various ray-tracing problems of astrophysics [ ATR , DSGGH , LE , M , DG1 , DG2 ] . We have proved that this problem is equivalent to the question of finding a formula for the so called triangular ratio metric of geometric function theory and also studied the problem in the case of quartic mirrors [ FHMV , FMV1 , FMV2 , HKV ] . Since the publication of our work [ FHMV ] , our solution has been implemented and validated in astrophysics applications in [ DG1 , DG2 ] and its performance compared to other methods. V.H. Almeida Jr., one of the authors of [ DG1 , DG2 ] , has sent us an email [ D ] where he pointed out the need for the solution of the problem in the case of the source at infinity. A few days after the receipt of his email we informed him about the solution of this problem and gave the solution in Theorem 3.1 below. The solution is given by a fourth degree polynomial equation which one can solve with symbolic computation. 2. Preliminaries For the reader’s benefit we recall here some of our earlier work [ FHMV , HKV ] . The triangular ratio metric s_{G} of a given domain G\subset{\mathbb{R}}^{2} is defined as follows for z_{1},z_{2}\in G [ HKV ] s_{G}(z_{1},z_{2})=\sup_{z\in\partial G}\frac{|z_{1}-z_{2}|}{|z_{1}-z|+|z-z_{2}|}\,. By compactness, this supremum is attained at some point w\in\partial G\,. If G is convex, it is simple to see that w is the point of contact of the boundary with an ellipse, with foci z_{1},z_{2}\,, contained in G\,. Now for the case G=\mathbb{B}^{2} and z_{1},z_{2}\in\mathbb{B}^{2}\, , if the extremal point is w\in\partial\mathbb{B}^{2}\,, then (2.1) s_{\mathbb{B}^{2}}(z_{1},z_{2})=\frac{|z_{1}-z_{2}|}{|z_{1}-w|+|z_{2}-w|}\,. This definition also makes sense for the domain \mathbb{R}^{2}\setminus\overline{\mathbb{B}}^{2}, we call the cases \mathbb{B}^{2} and \mathbb{R}^{2}\setminus\overline{\mathbb{B}}^{2} the interior and exterior problems, resp. The connection between the triangular ratio distance and the Ptolemy-Alhazen interior problem is now clear: if the unit circle is a reflecting mirror and z_{1} is the light source and z_{2} is the observer, then the point w is the point of reflection. We also see that an ellipse with foci z_{1} and z_{2} through the point w is internally tangent to the unit circle. 2.2 . Ellipse eccentricity. Let z_{1},z_{2}\in\mathbb{B}^{2} and w\in\partial\mathbb{B}^{2} be the extremal point of ( 2.1 ). Then the focal sum of the ellipse is c=|z_{1}-w|+|z_{2}-w|=|z_{1}-z_{2}|/s_{\mathbb{B}^{2}}(z_{1},z_{2})\,. Let \lambda,\sigma be the lengths of the major and minor semiaxes of the ellipse. Then \lambda=c/2\,,\quad\sigma=\frac{1}{2}\sqrt{c^{2}-|z_{1}-z_{2}|^{2}} and the eccentricity equals \sqrt{\frac{\lambda^{2}-\sigma^{2}}{\lambda^{2}}}=\sqrt{1-\left(\frac{\sigma}{\lambda}\right)^{2}}=s_{\mathbb{B}^{2}}(z_{1},z_{2})\,. In conclusion, s_{\mathbb{B}^{2}}(z_{1},z_{2}) is equal to the eccentricity of the maximal ellipse in {\mathbb{B}^{2}} with foci z_{1},z_{2}\,. Figure 1 . If the unit circle is a mirror, then a light signal from the source z_{1} to the observer at z_{2} is reflected at the point w. Moreover, s_{\mathbb{B}^{2}}(z_{1},z_{2}) is equal to the eccentricity of the maximal ellipse. The main result of [ FHMV ] , Theorem 2.3 , gives a polynomial equation of degree four that yields the reflection point on the unit circle. The solution of the quartic equation is given by Cardano’s formula which is available in standard symbolic computation software. We also study the Ptolemy-Alhazen exterior problem for z_{1},z_{2}\in\mathbb{R}^{2}\setminus\overline{\mathbb{B}}^{2}, and in this case the formula for the reflection point, if such a point exists, is the same as for the interior problem [ FHMV , Rmk 2.7] . Theorem 2.3 . For z_{1},z_{2}\in{\mathbb{B}}^{2}, the point w in ( 2.1 ) is given as a solution of the equation (2.4) \overline{z_{1}}\overline{z_{2}}w^{4}-(\overline{z_{1}}+\overline{z_{2}})w^{3}+(z_{1}+z_{2})w-z_{1}z_{2}=0\,. 2.5 . Remark. The roots of equation ( 2.4 ) have been studied in detail in [ FHMV ] . The root w at the extremal point for ( 2.1 ) is called the minimizing root. For instance, it is known that the minimizing root need not be unique and that some of the roots of ( 2.4 ) may be off \partial{\mathbb{B}^{2}}. A computer algorithm for finding the minimizing root and a formula for s_{\mathbb{B}^{2}}(z_{1},z_{2}) based on this root is given in [ FHMV ] . Corollary 2.6 . For z_{1},z_{2}\in\mathbb{B}^{2}, we have s_{\mathbb{B}^{2}}(z_{1},z_{2})=\frac{|z_{1}-z_{2}|}{|z_{1}-w|+|z_{2}-w|} where w\in\partial{\mathbb{B}^{2}} is the minimizing root of ( 2.4 ) . 2.7 . Testing and validation. The above method has been applied in astrophysics applications. Some of the comments read as follows. (1) From [ DG1 , p.7] : "Furthermore, we compared algorithms at 45^{o} to Fujimura et al.’s algorithm. We found Martín–Neira and Helm had a systematic discrepancy in terms of grazing angle and arc length, proportional to antenna heights. Fermat’s iterative algorithm presented the least accurate results. Therefore, we advise users to give preference to Fujimura et al.’s algorithm." (2) From [ DG2 , p.3304] : "Several formulations were previously tested and we found the model of Fujimura et al. (2019) to be more numerically stable (Almeida Junior and Geremia–Nievinski, 2023)." (3) From [ MBM ] : "Fujimura et al. (2019) provides a formulation that works well for the method of path–length minimization, but suffers numerical instability in cases where the radius of the sphere is very small relative to source or observer distance (e.g., in the one–finite case discussed below). Glaeser (1999) demonstrates the numerical root–finding approach (using an algorithm from Schwarze 1990) and notes that it is subject to "numerical instabilities that may lead to a considerable loss of accuracy" under certain circumstances." 3. Quartic equation Assume that light rays reach the unit circle from the right side, parallel to the real axis. Figure 2 . Light rays reach the unit circle from the right side. Let f=re^{i\theta} be the observation point, where we can assume r>1 and 0\leq\theta\leq\frac{\pi}{2} from symmetry. Let w be the reflection point and \widetilde{w} with \widetilde{w}=w+1 . Note that w and \widetilde{w} are points on a line parallel to the real axis. Figure 3 . The light ray reflects at the point w on the unit circle and reaches the point f . Theorem 3.1 . The reflection point w is given as a solution of the equation (3.2) re^{-i\theta}w^{4}-w^{3}+w-re^{i\theta}=0. Proof. From the reflection property, the following holds \measuredangle(0,w,\widetilde{w})=\measuredangle(f,w,0). Therefore, \displaystyle\arg\dfrac{\widetilde{w}-w}{0-w}=\arg\dfrac{0-w}{f-w} holds and \displaystyle\arg\dfrac{\widetilde{w}-w}{-w}-\arg\dfrac{-w}{f-w} \displaystyle=\arg\dfrac{w-\widetilde{w}}{w}\cdot\dfrac{w-f}{w} \displaystyle=\arg\dfrac{(w-(1+w))(w-f)}{w^{2}}=\arg\dfrac{f-w}{w^{2}}=0. The last equality implies that \dfrac{f-w}{w^{2}} is a real number. Since this complex conjugate is also real, we have \frac{f-w}{w^{2}}=\frac{\overline{f}-\overline{w}}{\overline{w}^{2}}. Hence, we have \overline{w}^{2}(re^{i\theta}-w)=w^{2}(re^{-i\theta}-\overline{w}). Since w is a point on the unit circle and satisfies w\overline{w}=1 , the above

Read the complete text