# GhostVision
# 🚧**UNDER CONSTRUCTION**🚧
[)](https://pypi.org/project/ghostvision/)
[](https://github.com/PINGEcosystem/GhostVision/commits)
[](https://github.com/PINGEcosystem/GhostVision/commits)
[](https://github.com/PINGEcosystem/GhostVision/blob/main/LICENSE)
Near-real time detection of derelict (ghost) crab pots with side-scan sonar.
## Overview
`GhostVision` is an open-source Python interface for automatically detecting and mapping ghost (derelict) crab pots from side-scan sonar imagery. `GhostVision` leverages [`Yolo`](https://docs.ultralytics.com/) models trained with [`Roboflow`](https://roboflow.com/). Detections are then georeferenced with [`PINGMapper`](https://github.com/CameronBodine/PINGMapper).
## Installation
### GPU (Fast Inference)
`GhostVision` is optimized for running inference (predictions) on the GPU. The processing environment is installed with `conda`. Any flavor of `conda` will do, but we recommend [`Miniforge`](https://conda-forge.org/download/). Follow the instructions below based on your OS.
#### Windows Only
Windows does not natively support inference on the GPU. A utility called [WSL](https://learn.microsoft.com/en-us/windows/wsl/install) (Windows Subsystem for Linux) needs to be installed in order to run inference on the GPU.
1. Install [WSL](https://learn.microsoft.com/en-us/windows/wsl/install) (Windows Subsystem for Linux) &
2. Open the command prompt by launching `Ubuntu` from the Windows Start menu.
#### Install `Miniforge`
3. In a command prompt, download [`Miniforge`](https://conda-forge.org/download/) with:
```
wget "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-$(uname)-$(uname -m).sh"
```
4. Install [`Miniforge`](https://conda-forge.org/download/) with:
```
bash Miniforge3-$(uname)-$(uname -m).sh
```
#### Install `GhostVision`
5. Install `PINGInstaller`:
```
pip install pinginstaller
```
6. Install `GhostVision`:
```
python -m pinginstaller ghostvision-gpu
```
### CPU (Slow Inference; Experimental)
An experimental version of `GhostVision` is available to test inference speeds on the CPU. This has been tested on Windows 11 only.
1. Install [`Miniforge`](https://conda-forge.org/download/).
2. Open the [`Miniforge`](https://conda-forge.org/download/) prompt.
3. Install `PINGInstaller`:
```
pip install pinginstaller
```
4. Install `GhostVision`.
```
python -m pinginstaller ghostvision
```
## Usage
1. Open the appropriate command prompt based on your installation above.
2. Launch `GhostVision`:
```
conda activate ghostvision
python -m ghostvision
```
3. Select desired parameters and click `Submit`.
## Download Custom `Roboflow` Object Detection Model
`GhostVision` includes `Roboflow` object detection models designed to detect crab pots from side-scan sonar imagery. You can train and use your own object detection model by downloading the model from `Roboflow` with the included utility.
1. Open the appropriate command prompt based on your installation above.
2. Launch the Roboflow model download utility:
```
conda activate ghostvision
python -m ghostvision rf-download
```
3. Supply your [Roboflow API Key](https://docs.roboflow.com/developer/authentication/find-your-roboflow-api-key).
4. Enter the project name (all lowercase).
5. Enter the project version.
The model will be downloaded and available to use.
## Acknowledgments
**Development Team:** [Cameron Bodine](https://github.com/CameronBodine), [Art Trembanis](https://www.udel.edu/academics/colleges/ceoe/departments/smsp/faculty/arthur-trembanis/), Kleio Baxevani, Onur Bagoren, Olivia Hines, Jared Wierzbicki, Ophelia Christoph, Catherine Hughes, Julia Greco.
Raw data
{
"_id": null,
"home_page": null,
"name": "ghostvision",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.6",
"maintainer_email": null,
"keywords": "pingmapper, sonar, ecology, remotesensing, sidescan, sidescan-sonar, aquatic, humminbird, lowrance, gis, oceanography, limnology, object-detection",
"author": "Cameron Bodine",
"author_email": "bodine.cs@gmail.email",
"download_url": "https://files.pythonhosted.org/packages/9a/d0/1881964e067fff40420d8316e5782d1e8fe19a288e20d6df5520b8989567/ghostvision-1.0.0b1.tar.gz",
"platform": null,
"description": "# GhostVision\r\n\r\n# \u00f0\u0178\u0161\u00a7**UNDER CONSTRUCTION**\u00f0\u0178\u0161\u00a7\r\n\r\n[)](https://pypi.org/project/ghostvision/)\r\n[](https://github.com/PINGEcosystem/GhostVision/commits)\r\n[](https://github.com/PINGEcosystem/GhostVision/commits)\r\n[](https://github.com/PINGEcosystem/GhostVision/blob/main/LICENSE)\r\n\r\nNear-real time detection of derelict (ghost) crab pots with side-scan sonar.\r\n\r\n\r\n\r\n## Overview\r\n\r\n`GhostVision` is an open-source Python interface for automatically detecting and mapping ghost (derelict) crab pots from side-scan sonar imagery. `GhostVision` leverages [`Yolo`](https://docs.ultralytics.com/) models trained with [`Roboflow`](https://roboflow.com/). Detections are then georeferenced with [`PINGMapper`](https://github.com/CameronBodine/PINGMapper).\r\n\r\n## Installation\r\n\r\n### GPU (Fast Inference)\r\n\r\n`GhostVision` is optimized for running inference (predictions) on the GPU. The processing environment is installed with `conda`. Any flavor of `conda` will do, but we recommend [`Miniforge`](https://conda-forge.org/download/). Follow the instructions below based on your OS.\r\n\r\n#### Windows Only\r\nWindows does not natively support inference on the GPU. A utility called [WSL](https://learn.microsoft.com/en-us/windows/wsl/install) (Windows Subsystem for Linux) needs to be installed in order to run inference on the GPU.\r\n\r\n1. Install [WSL](https://learn.microsoft.com/en-us/windows/wsl/install) (Windows Subsystem for Linux) & \r\n2. Open the command prompt by launching `Ubuntu` from the Windows Start menu.\r\n\r\n#### Install `Miniforge`\r\n\r\n3. In a command prompt, download [`Miniforge`](https://conda-forge.org/download/) with:\r\n ```\r\n wget \"https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-$(uname)-$(uname -m).sh\"\r\n ```\r\n4. Install [`Miniforge`](https://conda-forge.org/download/) with:\r\n ```\r\n bash Miniforge3-$(uname)-$(uname -m).sh\r\n ```\r\n\r\n#### Install `GhostVision`\r\n\r\n5. Install `PINGInstaller`:\r\n ```\r\n pip install pinginstaller\r\n ```\r\n6. Install `GhostVision`:\r\n ```\r\n python -m pinginstaller ghostvision-gpu\r\n ```\r\n\r\n### CPU (Slow Inference; Experimental)\r\nAn experimental version of `GhostVision` is available to test inference speeds on the CPU. This has been tested on Windows 11 only.\r\n\r\n1. Install [`Miniforge`](https://conda-forge.org/download/).\r\n2. Open the [`Miniforge`](https://conda-forge.org/download/) prompt.\r\n3. Install `PINGInstaller`:\r\n ```\r\n pip install pinginstaller\r\n ```\r\n4. Install `GhostVision`.\r\n ```\r\n python -m pinginstaller ghostvision\r\n ```\r\n\r\n## Usage\r\n\r\n1. Open the appropriate command prompt based on your installation above.\r\n2. Launch `GhostVision`:\r\n ```\r\n conda activate ghostvision\r\n python -m ghostvision\r\n ```\r\n3. Select desired parameters and click `Submit`.\r\n\r\n## Download Custom `Roboflow` Object Detection Model\r\n\r\n`GhostVision` includes `Roboflow` object detection models designed to detect crab pots from side-scan sonar imagery. You can train and use your own object detection model by downloading the model from `Roboflow` with the included utility.\r\n\r\n1. Open the appropriate command prompt based on your installation above.\r\n2. Launch the Roboflow model download utility:\r\n ```\r\n conda activate ghostvision\r\n python -m ghostvision rf-download\r\n ```\r\n3. Supply your [Roboflow API Key](https://docs.roboflow.com/developer/authentication/find-your-roboflow-api-key).\r\n4. Enter the project name (all lowercase).\r\n5. Enter the project version.\r\n\r\nThe model will be downloaded and available to use.\r\n\r\n## Acknowledgments\r\n\r\n**Development Team:** [Cameron Bodine](https://github.com/CameronBodine), [Art Trembanis](https://www.udel.edu/academics/colleges/ceoe/departments/smsp/faculty/arthur-trembanis/), Kleio Baxevani, Onur Bagoren, Olivia Hines, Jared Wierzbicki, Ophelia Christoph, Catherine Hughes, Julia Greco.\r\n\r\n\r\n",
"bugtrack_url": null,
"license": null,
"summary": "Near-real time detection of derelict (ghost) crab pots with side-scan sonar.",
"version": "1.0.0b1",
"project_urls": {
"GitHub": "https://github.com/PINGEcosystem/GhostVision",
"Issues": "https://github.com/PINGEcosystem/GhostVision/issues"
},
"split_keywords": [
"pingmapper",
" sonar",
" ecology",
" remotesensing",
" sidescan",
" sidescan-sonar",
" aquatic",
" humminbird",
" lowrance",
" gis",
" oceanography",
" limnology",
" object-detection"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "6a8971ce1809bacea686baeab8753990398c272caef39a9c88411bf7f2838d7f",
"md5": "273e4dfeab71d8e18189e44b9a179944",
"sha256": "e9a461ecb3eb3a93fe920d61be23c465b44ec9da95889cf31528a2ecc8ce02ee"
},
"downloads": -1,
"filename": "ghostvision-1.0.0b1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "273e4dfeab71d8e18189e44b9a179944",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.6",
"size": 15914,
"upload_time": "2025-10-08T12:47:03",
"upload_time_iso_8601": "2025-10-08T12:47:03.596324Z",
"url": "https://files.pythonhosted.org/packages/6a/89/71ce1809bacea686baeab8753990398c272caef39a9c88411bf7f2838d7f/ghostvision-1.0.0b1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "9ad01881964e067fff40420d8316e5782d1e8fe19a288e20d6df5520b8989567",
"md5": "952ae62d93e8719d7c2b2836cf870cd4",
"sha256": "391f89df50c2dff7c22fe03ddc3da0a6cc054f834c89c2c99aad09cf8f880c37"
},
"downloads": -1,
"filename": "ghostvision-1.0.0b1.tar.gz",
"has_sig": false,
"md5_digest": "952ae62d93e8719d7c2b2836cf870cd4",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.6",
"size": 16573,
"upload_time": "2025-10-08T12:47:05",
"upload_time_iso_8601": "2025-10-08T12:47:05.145065Z",
"url": "https://files.pythonhosted.org/packages/9a/d0/1881964e067fff40420d8316e5782d1e8fe19a288e20d6df5520b8989567/ghostvision-1.0.0b1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-10-08 12:47:05",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "PINGEcosystem",
"github_project": "GhostVision",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "ghostvision"
}
JSON
