6 releases

0.2.1 Apr 3, 2023
0.2.0 Mar 16, 2023
0.1.6 Mar 15, 2023
0.1.5 May 11, 2022
0.1.1 Nov 24, 2021

#109 in Science

31 downloads per month

MIT license

29KB
256 lines

PyPlanetarium

Python bindings for Planetarium sub-pixel precision light spot rendering library for astronomy and video tracking applications.

Sample image

Sample

Example usage

from pyplanetarium import Canvas, SpotShape

# Draw on a square 256x256 pixel canvas.
c = Canvas.new(256, 256)

# Define a round spot shape with diffraction radius of 2.5 pixels.
shape = SpotShape().scale(2.5)

# Add some spots at random positions with varying shape size
# and peak intensity.
spot1 = c.add_spot((100.3, 130.8), shape, 0.5)
spot2 = c.add_spot((80.6, 200.2), shape.scale(0.5), 0.9)

# Note: Out of range position coordinates and peak intensities are fine.
#       The resulting spot image is clipped into the canvas rectangle.
#       Peak intensity > 1.0 leads to saturation to the maximum pixel value.
spot3 = c.add_spot((256.1, 3.5), shape.scale(10.0), 1.1)

# Set the canvas background pixel value.
c.set_background(int(0.05 * Canvas.PIXEL_MAX))

# Clear the canvas and paint the light spots.
c.draw()

# Get image dimensions.
image_width, image_height = c.dimensions()

Light spot parameters adjustment

Some of the light spot parameters like coordinates and peak intensity can be adjusted after the spot has been added to the canvas.

The spot position coordinates can be changed by adding an offset vector and the peak intensity can be adjusted by multiplying with a spot illumination factor.

It is possible to define a custom world coordinates to canvas coordinates transformation, which affects all light spots.

from pyplanetarium import Canvas, SpotShape, Transform

# Draw on a square 256x256 pixel canvas.
c = Canvas.new(256, 256)

# Define an elliptic spot shape with diffraction radii of 2.5 x 1.5 pixels
# rotated by 45 degrees counter-clockwise.
shape1 = SpotShape().stretch(2.5, 1.5).rotate(45.0)

# Define an elliptic spot shape by a 2x2 linear transform matrix.
shape2 = SpotShape([[2.0, -0.5], [1.5, 3.0]])

# Add some spots at random positions with varying shape size
# and peak intensity.
spot1 = c.add_spot((100.3, 130.8), shape1, 0.5)
spot2 = c.add_spot((80.6, 200.2), shape2, 0.9)

# Shift the rendered spot positions by applying the relative offset vectors.
# The intrinsic spot position coordinates are immutable.
c.set_spot_offset(spot1, (-34.2, 12.6))
c.set_spot_offset(spot2, (114.2, -73.3))

# Adjust the rendered spot peak intensity by applying the spot illumination factors.
# The intrinsic spot intensities are immutable.
c.set_spot_illumination(spot1, 1.2)
c.set_spot_illumination(spot2, 0.7)

# Query the resulting spot coordinates on the canvas.
pos1 = c.spot_position(spot1)
pos2 = c.spot_position(spot2)

# Query the resulting peak spot intensities.
int1 = c.spot_intensity(spot1)
int2 = c.spot_intensity(spot2)

# Apply a custom world coordinates to canvas coordinates transformation.
c.set_view_transform(Transform().translate((13.7, -20.3)))

# Query the resulting spot coordinates on the canvas after
# the view coordinate transformation.
pos1x = c.spot_position(spot1)
pos2x = c.spot_position(spot2)

Canvas image export

The Canvas object supports image export to RAW and PNG file formats. Both 8-bit and 16-bit PNG sample formats are supported. Export to PNG formats requires the default png feature to be enabled.

Example image export code

from pyplanetarium import Canvas, ImageFormat

c = Canvas.new(256, 256)

c.set_background(1000)
c.clear()

# Export to a 8-bit gamma-compressed grayscale RAW image.
raw_8bpp_bytes = c.export_image(ImageFormat.RawGamma8Bpp)

# Export to a 10-bit linear light grayscale little-endian RAW image.
raw_10bpp_bytes = c.export_image(ImageFormat.RawLinear10BppLE)

# Export to a 12-bit gamma-compressed grayscale little-endian RAW image.
raw_12bpp_bytes = c.export_image(ImageFormat.RawLinear12BppLE)

# Export to a 8-bit gamma-compressed grayscale PNG image.
png_8bpp_bytes = c.export_image(ImageFormat.PngGamma8Bpp)

# Export to a 16-bit linear light grayscale PNG image.
png_16bpp_bytes = c.export_image(ImageFormat.PngLinear16Bpp)

Window image export

The Canvas object additionally supports windowed image export.

A single rectangular window represents a region of interest (ROI) on the canvas image. Window rectangle coordinates are represented by the public Window structure.

Example window image export code

from pyplanetarium import Canvas, ImageFormat, Window

c = Canvas.new(256, 256)

# Create a 64x32 pixels window with origin at (90, 120).
wnd = Window.new(64, 32).at(90, 120)

fmt = ImageFormat.RawGamma8Bpp

# Export to the canvas window image bytes.
raw_window_bytes = c.export_window_image(wnd, fmt)

Subsampled image export

The Canvas object additionally supports subsampled image export with independent row and column subsampling factors.

Only whole canvas images can be exported with subsampling.

Example subsampled image export code

from pyplanetarium import Canvas, ImageFormat

c = Canvas.new(256, 256)

fmt = ImageFormat.RawLinear10BppLE

# Column (X) and row (Y) subsampling factors
factors = (2, 2)

# Export to the subsampled canvas image bytes.
raw_sub_bytes = c.export_subsampled_image(factors, fmt)

Dependencies

~4–10MB
~79K SLoC