Tracing Disks

A 2D ray tracer for computing scattering from many small disks (or a single large disk). Rays reflect/refract using Fresnel equations, accumulate energy, and produce an angular distribution. Includes a plot mode that records geometry + ray paths so we can generate an animation.

⚙️ Build

Linux

g++ -O3 -std=c++17 -fopenmp tracedisks.cpp -o trace

🚀 Run

1) Single large disk

Produces a .bin distribution file.

./trace single

2) Many small disks (averaged over experiments)

./trace multiple

3) Plot mode (record one random configuration for visualization)

Writes circles.csv and rays.json for the Python animator.

./trace multiple --plot

Common flags

Flag	Description	Default
--num_paths	Rays per experiment	1000000
--num_circles	Maximum number of small disks	100
--radius_small	Radius of each small disk	1.0
--n1, --n2	Refractive indices outside/inside	1.00 / 1.50
--sigma_a	Absorption coefficient	0.0
--num_exp	Experiments to average (multiple mode)	100
--epsilon	Minimum gap between disks when packing	0.1
--max_bounces	Max per-ray interactions	100
--num_bins	Angular histogram bins	100
--plot	Enable recording for visualization	—

🎨 Visualize

Use plot.ipynb to convert circles.csv + rays.json into a video, or visualize the scattering distributions.

📂 Output files

• single_sigma_*_n1_*_n2_*.bin, multiple_avg_sigma_*_n1_*_n2_*_radius_*.bin: binary arrays of length num_bins, the angular energy distribution.
• circles.csv: center (cx, cy) and radius r for each disk (first row = big boundary radius at (0,0)).
• rays.json: one ray per line, e.g. { "points": [[x,y],...], "out_angle": ..., "energy": ... }.

🧠 How it works

• Disks are packed randomly with a grid-accelerated rejection sampler (no overlaps).
• Rays start at the left; at each hit: compute normal, Fresnel reflectance; reflect or refract; attenuate via absorption.
• Outgoing rays’ angles and energies are binned into an angular histogram.
• In plot mode, a single configuration’s disks and a subset of rays are recorded for animation.