gradoptics.optics.sensor

class gradoptics.optics.sensor.Sensor(resolution=(9600, 9600), pixel_size=(3.76e-06, 3.76e-06), position=(- 0.0575, 0, 0), poisson_noise_mean=2, quantum_efficiency=0.8, viewing_direction=[1.0, 0.0, 0.0], up=[0.0, 0.0, 1.0], psfs={}, psf_ratio=1)

Bases: gradoptics.optics.base_optics.BaseOptics

Models a sensor.

__init__(resolution=(9600, 9600), pixel_size=(3.76e-06, 3.76e-06), position=(- 0.0575, 0, 0), poisson_noise_mean=2, quantum_efficiency=0.8, viewing_direction=[1.0, 0.0, 0.0], up=[0.0, 0.0, 1.0], psfs={}, psf_ratio=1)
Parameters

  • resolution – Image processing convention: origin in the upper left corner, horizontal x axis and vertical y axis (tuple)

  • pixel_size – Image processing convention: origin in the upper left corner, horizontal x axis and vertical y axis (tuple)

  • position – Position of the lens (list)

  • poisson_noise_mean – Mean readout noise (float)

  • quantum_efficiency – Quantum efficiency (float)

  • viewing_direction – Viewing direction of the sensor (list)

  • up – A vector that orients the sensor with respect to the viewing direction (list) For example, if up=[0., 0., 1.] the top of the camera will point upwards. If up=[0., 0., -1.], the top of the camera will point downwards.

  • psfs – Depth and position-dependant point spread function (dict)

  • psf_ratio – Ratio between the psf resolution and the sensor resolution (int)

get_ray_intersection(incident_rays)

Computes the times t at which the incident rays will intersect the object.

Parameters

incident_rays – The incident rays (Rays)

Returns

Times t (torch.tensor)

intersect(incident_rays, t, do_pixelize=True, quantum_efficiency=True)

Returns the rays reflected or refracted by the optical element.

Parameters

  • incident_rays – The incident rays (Rays)

  • t – The times at which the incident rays will intersect the optical element (torch.tensor)

Returns

(tuple) - Reflected or refracted rays (Rays) - A boolean mask that indicates which incident rays were reflected or refracted ― some rays may not be

reflected or refracted for example, due to total internal reflection, which is not modelled at this point (torch.tensor)

pixelize(depth_id, hit_positions, luminosities, quantum_efficiency=True)

Accumulates photons at the pixels hit by some rays at the positions hit_positions

Parameters

  • depth_id – index of the depth (encoded in the PSF) from where the rays come from (int)

  • hit_positions – The positions at which the rays hit the sensor (torch.tensor)

  • luminosities – The luminosities of the rays (torch.tensor)

  • quantum_efficiency – Whether to use quantum efficiency or no (bool)

readout(add_poisson_noise=True, destructive_readout=True)

Readouts the sensor

Parameters

  • add_poisson_noise – Whether to add poisson noise or no (bool)

  • destructive_readout – Whether the accumulated photons should be destroyed or no (bool)

Returns

The produced image (torch.tensor)

plot(ax, s=0.1, color='grey', resolution_=100)

Plots the optical element on the provided axes.

Parameters

ax – 3d axes (mpl_toolkits.mplot3d.axes3d.Axes3D)

sample_points_on_sensor(nb_points, device='cpu')

Sample points uniformly on the sensor

Parameters

  • nb_points – Number of points to sample (int)

  • device – The desired device of returned tensor (str). Default is 'cpu'

Returns

Sampled points (torch.tensor)

sample_on_pixels(samples_per_pixel, idx, idy, device='cpu')

Sample data on the pixels whose id are provided

Parameters

  • samples_per_pixel – Number of samples per pixel (int)

  • idx – Pixel indices along the horizontal axis, assuming a coordinate system centered in the center of the image(torch.tensor)

  • idy – Pixel indices along the vertical axis, assuming a coordinate system centered in the center of the image(torch.tensor)

  • device – The desired device of returned tensor (str). Default is 'cpu'

Returns

(tuple) Sampled points (torch.tensor) and p(A) (float)