---
name: "celeste_standalone"
section: "1"
platform: "linux"
package: "hugin"
version: "2025.0.1-5"
release: "Arch Linux 2026-11-01"
platforms: ["linux"]
url: "https://new.carta.tech/celeste_standalone"
---

# celeste\_standalone(1)

Cloud identification

## NAME

celeste\_standalone - Cloud identification

## SYNOPSIS

**celeste\_standalone** \[options\] *image1* *image2* *\[..\]*

## DESCRIPTION

Celeste has been trained using Support vector machine techniques to identify clouds in photos and remove control points from these areas.  celeste\_standalone is a command-line tool with all the same functionality as Celeste in hugin.

Simple usage is to just 'clean' an existing project file:

```
 celeste_standalone -i project.pto -o project.pto
```

## OPTIONS

- **-i** \<filename\>

  Input Hugin PTO file. Control points over SVM threshold will be removed before being written to the output file. If -m is set to 1, images in the file will be also be masked.

- **-o** \<filename\>

  Output Hugin PTO file. Default: '\<filename\>\_celeste.pto'

- **-d** \<filename\>

  SVM model file. Default: 'data/celeste.model'

- **-s** \<int\>

  Maximum dimension for re-sized image prior to processing. A higher value will increase the resolution of the mask but is significantly slower. Default: 800

- **-t** \<float\>

  SVM threshold. Raise this value to remove fewer control points, lower it to remove more. Range 0 to 1. Default: 0.5

- **-m** \<1\|0\>

  Create masks when processing Hugin PTO file. Default: 0

- **-f** \<string\>

  Mask file format. Options are PNG, JPEG, BMP, GIF and TIFF.  Default: PNG

- **-r** \<1\|0\>

  Filter radius. 0 = large (more accurate), 1 = small (higher resolution mask, slower, less accurate). Default: 0

- **-h**

  Print usage.

## AUTHORS

Written by Tim Nugent.