nickyspatial 1.1.1

An open-source object-based image analysis library for remote sensing.

NickySpatial

An open-source object-based image analysis library for remote sensing.

Category	Badge
Build
Code Quality
Documentation
Package Info
Community

[!WARNING] This project is under active development and lot of its functionality is still in my head yet to code.

Find Demo Frontend Here : https://nickyspatial-gpoqz.ondigitalocean.app/

Description

NickySpatial is a Python package that provides object-based image analysis (OBIA) functionality similar to commercial software like eCognition. It allows users to segment geospatial imagery into meaningful objects, calculate statistics, and apply rule-based classification.

Project Structure

nickyspatial/
├── __init__.py
├── io/
│   ├── __init__.py
│   ├── raster.py       # Raster I/O
│   └── vector.py       # Vector I/O
├── core/
│   ├── __init__.py
│   ├── layer.py        # Layer class and management
│   ├── segmentation.py # Segmentation algorithms
│   └── rules.py        # Rule engine
|   └── classifier.py   # Classification algorithms
├── stats/
│   ├── __init__.py
│   ├── basic.py        # Basic statistics (min, max, mean, etc.)
│   ├── spatial.py      # Spatial statistics (area, perimeter, etc.)
│   └── spectral.py     # Spectral indices (NDVI, etc.)
├── filters/
│   ├── __init__.py
│   ├── spatial.py      # Spatial filters (smoothing, merging)
│   └── spectral.py     # Spectral filters (band math)
├── viz/
│   ├── __init__.py
│   ├── maps.py         # Map visualization
│   └── charts.py       # Statistical charts
└── utils/
    ├── __init__.py
    └── helpers.py      # Helper functions

Installation

pip install nickyspatial

Quick Start

import nickyspatial as ns

# Load a raster image
layer = ns.read_raster("path/to/your/satellite_image.tif")

# Perform segmentation
segments = ns.SlicSegmentation(scale=20, compactness=0.5)
segmented_layer = segments.execute(layer.raster, layer.transform, layer.crs)

# Calculate statistics
ns.attach_basic_stats(segmented_layer, "mean_intensity")
ns.attach_area_stats(segmented_layer)
ns.attach_spectral_indices(segmented_layer, indices=["ndvi"])

# Apply rules for classification
ruleset = ns.RuleSet(name="Land_Cover")
ruleset.add_rule(name="vegetation", condition="ndvi > 0.3 & area_units > 100")
classified_layer = ruleset.execute(segmented_layer, result_field="classification")

# Visualize results
ns.plot_layer_interactive(classified_layer, "classification")

Documentation

For detailed documentation and examples, see the documentation website. Deepwiki also provides documentation quite detailed , Do check it out if you like

Examples

Check out our comprehensive examples:

• Basic Usage: Complete workflow from loading data to visualization
• Supervised Classification: Machine learning-based classification with traditional algorithms
• CNN Classification: Deep learning approach using TensorFlow/Keras
• Interactive Demo: Web-based interface for exploring features

Key Features Demonstrated:

• Image Segmentation: SLIC algorithm for object-based analysis
• Statistical Analysis: Calculate spatial, spectral, and basic statistics
• Classification: Both traditional ML and deep learning approaches
• Rule Engine: Apply custom rules for object classification
• Visualization: Interactive maps and statistical charts
• I/O Operations: Read/write raster and vector formats

Contributing

Contributions are welcome! Follow dev setup guide & Please feel free to submit a Pull Request.

Acknowledgments

• Inspired by the functionality of eCognition and other OBIA methodologies
• Built on top of powerful open-source libraries like numpy, rasterio, scikit-image, and GeoPandas
• Nicky : Nicky is my belated dog and I named this package in his memory!

  

Contributors