We present xBD, a new, large-scale dataset for the advancement of change detection and building damage assessment for humanitarian assistance and disaster recovery research. Natural disaster response requires an accurate understanding of damaged buildings in an affected region. Current response strategies require in-person damage assessments within 24-48 hours of a disaster. Massive potential exists for using aerial imagery combined with computer vision algorithms to assess damage and reduce the potential danger to human life. In collaboration with multiple disaster response agencies, xBD provides pre-and post-event satellite imagery across a variety of disaster events with building polygons, ordinal labels of damage level, and corresponding satellite metadata. Furthermore, the dataset contains bounding boxes and labels for environmental factors such as fire, water, and smoke. xBD is the largest building damage assessment dataset to date, containing 850,736 building annotations across 45,362 km 2 of imagery.

Unsustainable fishing practices worldwide pose a major threat to marine resources and ecosystems. Identifying vessels that evade monitoring systems-known as "dark vessels"-is key to managing and securing the health of marine environments. With the rise of satellite-based synthetic aperture radar (SAR) imaging and modern machine learning (ML), it is now possible to automate detection of dark vessels day or night, under all-weather conditions. SAR images, however, require domainspecific treatment and is not widely accessible to the ML community. Moreover, the objects (vessels) are small and sparse, challenging traditional computer vision approaches. We present the largest labeled dataset for training ML models to detect and characterize vessels from SAR. xView3-SAR consists of nearly 1,000 analysisready SAR images from the Sentinel-1 mission that are, on average, 29,400-by-24,400 pixels each. The images are annotated using a combination of automated and manual analysis. Co-located bathymetry and wind state rasters accompany every SAR image. We provide an overview of the results from the xView3 Computer Vision Challenge, an international competition using xView3-SAR for ship detection and characterization at large scale. We release the data (https://iuu.xview.us/) and code (https://github.com/DIUx-xView) to support ongoing development and evaluation of ML approaches for this important application. l IntroductionRecent advances in remote sensing technology have allowed fishing activity to be tracked across the globe via the Automatic Identification System (AIS) which can broadcast vessels' location [12]. Use of AIS, however, varies by region and fleet; not all vessels are required to carry AIS [29]; some turn their AIS off to engage in illicit activities [23]. This unknown number of non-broadcasting vessels that evade conventional monitoring systems-referred to as "dark" vessels-greatly limits our ability to manage marine resources. Illegal, unreported, and unregulated (IUU) fishing comprises over 20% of all catch around the world [1]. In recent years, the largest IUU fishing offenses were perpetrated by fleets that mostly did not use AIS [23], costing legitimate fishers and governments billions of dollars while also damaging critical ecological systems.Satellite imagery provides an alternative means of sensing dark vessels. Common electro-optical (EO) satellites, however, are limited by cloud coverage and low light conditions. Synthetic aperture radar (SAR) satellites, on the other hand, are able to image in all weather conditions and at nighttime. The European Space Agency (ESA) Sentinel-1 radar satellites cover all coastal waters around the world approximately every six days, offering open access to the full SAR archive. Despite its availability,

Deep learning tasks are often complicated and require a variety of components working together efficiently to perform well. Due to the often large scale of these tasks, there is a necessity to iterate quickly in order to attempt a variety of methods and to find and fix bugs. While participating in IARPA's Functional Map of the World challenge, we identified challenges along the entire deep learning pipeline and found various solutions to these challenges. In this paper, we present the performance, engineering, and deep learning considerations with processing and modeling data, as well as underlying infrastructure considerations that support large-scale deep learning tasks. We also discuss insights and observations with regard to satellite imagery and deep learning for image classification.

No abstract