# Astronomy PSF
**URL:** https://discourse.julialang.org/t/astronomy-psf/129430
**Category:** Astro/Space
**Tags:** question
**Created:** [May 29, 2025, 9:08am UTC](https://discourse.julialang.org/t/astronomy-psf/129430 "2025-05-29T09:08:31Z")
**Posts on this page:** 3
**Page:** 1
### Author:  [@Cyan](https://discourse.julialang.org/u/Cyan)
#### Post date: [May 29, 2025, 9:08am UTC](https://discourse.julialang.org/t/astronomy-psf/129430/1 "2025-05-29T09:08:31Z")
If I have a PSF and a unconvolve image, Does there exist a expert PSF package to convolve the image?
---
### Author:  [@cgarling](https://discourse.julialang.org/u/cgarling)
#### Post date: [June 5, 2025, 2:04pm UTC](https://discourse.julialang.org/t/astronomy-psf/129430/2 "2025-06-05T14:04:35Z")
If you can load your PSF and your image as matrices, then you can compute the convolution with [ImageFiltering.jl](https://juliaimages.org/ImageFiltering.jl/stable/kernels/#Correlation,-not-convolution). More information is given in the docs, but a short demo is given below
```julia
using ImageFiltering: reflect, imfilter, centered
psf = centered(rand(5,5)) # Use `centered` to properly configure matrix indices
psf ./= sum(psf) # Normalize psf to unity sum
# ImageFiltering.jl performs image correlation, not convolution, so we reflect
# the PSF before passing into `imfilter`
kernel = reflect(psf)
image = rand(100, 100)
conv_image = imfilter(image, kernel)
# `imfilter` supports many other options here (e.g., boundary conditions), see docs
```
---
### Author:  [@icweaver](https://discourse.julialang.org/u/icweaver)
#### Post date: [June 5, 2025, 4:02pm UTC](https://discourse.julialang.org/t/astronomy-psf/129430/3 "2025-06-05T16:02:52Z")
And here’s another example [with doggies](https://computationalthinking.mit.edu/Fall24/images_abstractions/transforming_images/#7489a570-74a3-11eb-1d0b-09d41604ffe1)!