Hello GeoLens

Script: 0_hello_geolens.py

The fastest end-to-end tour of the geometric-lens workflow: load a lens, run the classical optical analyses, and render an image through it with both ray tracing and PSF-map convolution.

What it demonstrates

Loading a GeoLens from JSON (Zemax .zmx and Code V are also supported).
The classical evaluation plots: layout, spot diagram, MTF, distortion, and relative illumination (vignetting).
Two rendering paths: per-ray ray_tracing and spatially-varying psf_map.

Run

python 0_hello_geolens.py

Key code

from deeplens import GeoLens

lens = GeoLens(filename="./datasets/lenses/cellphone/cellphone80deg.json")

lens.draw_layout(filename="./lens.png")
lens.analysis_spot()
lens.draw_spot_radial(save_name="./lens_spot.png")
lens.draw_mtf(depth_list=[lens.obj_depth], save_name="./lens_mtf.png")
lens.draw_distortion_radial(save_name="./lens_distortion.png")
lens.draw_vignetting(filename="./lens_vignetting.png", depth=lens.obj_depth)

# Image simulation at the sensor resolution
lens.set_sensor_res((3000, 2000))
img_ray = lens.render(img, depth=DEPTH, method="ray_tracing", spp=8)
img_psf = lens.render(img, depth=DEPTH, method="psf_map", psf_grid=(30, 20))