ArqPy 1.2.1

ArqPy 1.2.1 expands the toolbox beyond its core preprocessing workflow by incorporating AI-assisted tools for pansharpening and crop-mark exploration. The release remains focused on the reproducible processing of very-high-resolution satellite imagery for archaeological applications.

Main changes

• Added Z-PNN deep-learning pansharpening through pansharpening_cnn.bat.
• Added SAM 3 text-guided segmentation through sam3_detection.bat.
• Improved MAE analysis with selectable input bands, overlapping tiles, raised-Hann blending, optional vector clipping, and raster statistics.
• Added tools for extracting and inspecting MAE values at digitized crop-mark locations.
• Added detailed user guides for every toolbox operation.
• Improved graphical interfaces, input validation, output handling, and application branding.
• Reduced and updated the bundled Conda environments to keep the release below 2 GB.

Available operations

The bundled operations include:

1. Atmospheric correction
2. Conventional pansharpening
3. Principal component analysis
4. Spectral-index generation
5. High-pass filtering
6. MAE-based saliency analysis
7. Z-PNN deep-learning pansharpening

SAM 3 text-guided segmentation is available as an optional externally installed operation.

Sensor support

The core preprocessing workflow supports WorldView-3 and WorldView Legion-06 imagery. High-pass filtering and MAE analysis can process other GeoTIFF imagery. The Z-PNN module supports WorldView-2, WorldView-3, and GeoEye-1 inputs.

Installation

The release includes the environments required for preprocessing, MAE analysis, and Z-PNN pansharpening. Before running ArqPy, decompress env.7z, env_mae.7z, and env_pnn.7z inside the application directory.

The resulting structure must contain env, env_mae, and env_pnn directly beside the batch files. Avoid nested paths such as env\env or env_pnn\env_pnn.

SAM 3 is the only component requiring a separate environment and additional configuration. It also requires access to its pretrained checkpoint and a CUDA-compatible GPU. Consult the included documentation for setup instructions.

Important note

MAE and SAM 3 outputs support preliminary archaeological exploration. They identify salient or crop-mark-like patterns but cannot determine their archaeological origin. Expert interpretation and field validation remain necessary.

ArcPy Toolbox-1.0.1

A self-contained ArcPy toolbox for archaeological remote sensing, offering ready-to-run workflows for crop mark detection using spectral processing and Masked Autoencoder analysis on Windows 11.