YOLO-MOD Trained Models for Object Detection in Remote Sensing Imagery (DOTANA and ShipRSImageNet)

This repository provides trained YOLO-based object detection models used in the YOLO-MOD QGIS plugin. The models enable multi-class object detection in very high-resolution (VHR) optical remote sensing imagery and were developed to support reproducible experiments presented in the associated SoftwareX publication. The models are trained on two datasets: DOTANA (aircraft, helicopters, airports, storage tanks) ShipRSImageNet (civilian and military ships) The repository includes models in both PyTorch (.pt) and ONNX (.onnx) formats, allowing flexible deployment depending on user requirements and hardware configuration. Each model is accompanied by metadata including architecture type, training dataset, input resolution, and evaluation metrics (mAP50 and mAP50–95). These models are intended for use within the YOLO-MOD plugin but can also be applied independently in other object detection workflows. The YOLO-MOD plugin and methodology are presented in the accompanying SoftwareX publication: Ciecholewski, M., Strzelecki, M., & Farelnik, M. (2026).YOLO-MOD: YOLO-based multi-category object detection plugin for maritime, aviation, and infrastructure objects in optical remote sensing images in QGIS.SoftwareX, 34, 102721.https://doi.org/10.1016/j.softx.2026.102721