The paper associated with the dataset file is titled "Thunder: A Large-Scale EROS Dataset for Remote Sensing Image Change Detection" , published in 2024.
The paper introduces a baseline model often referred to as (or similar architecture variations), which utilizes a dual-stream encoder-decoder structure designed to handle the multi-scale variations and noise common in large-scale remote sensing imagery. The "Thunder" naming convention reflects the goal of achieving high-speed and high-accuracy processing for massive geospatial datasets. Typical Use Case If you have downloaded this .rar file, it likely contains: EROS_-_Thunder.rar
: Pre-change and post-change satellite images. The paper associated with the dataset file is
: Binary masks where white pixels (255) represent change and black pixels (0) represent no change. Typical Use Case If you have downloaded this
: It is one of the largest datasets for change detection, containing thousands of high-resolution (0.5m to 2.0m) image pairs.
: It specifically targets the "Easy-to-Retrieve and Open-Source" (EROS) philosophy to make high-quality change detection data accessible to the broader research community. Technical Contribution