CrowdAI’s Boat & Ship Detection model uses cutting-edge deep learning and image segmentation techniques to precisely locate and label boats and ships on Maxar satellite imagery. Detections are output as both a geoTIFF and polygons, providing denser, more precise labels than traditional bounding boxes. This model is globally flexible in line with Maxar’s coverage.
Create a Modzy account to get started →
72% F1 Score
99% Pixelwise Accuracy
The Boat & Ship Detection model accuracy scores demonstrate that this model correctly detects and labels vessels approximately 75% of the time, when accounting for both false positives and false negatives. Known objects that sometimes cause false positives include: floating docks far from shore; large buoys, especially those with a superstructure.
F1 Score is the harmonic mean of the precision and recall, with best value of 1. It measures the balance between the two metrics.
Pixelwise accuracy measures the percent of pixels in a predicted area that are classified or predicted correctly.
A higher precision score indicates that the majority of labels predicted by the model for different classes are accurate.
A higher recall score indicates that the model finds and predicts correct labels for the majority of the classes it is supposed to find.
Unlike traditional computer vision techniques, the Boat & Ship Detection model is powered by CrowdAI’s custom-built neural network for image segmentation. By classifying each pixel in the image, the model can provide precise vessel footprints, which are then converted into smoothed polygons as an output. This results in tighter detection profiles for each vessel—and thus better measurements of hull length and width—without sacrificing speed or accuracy.
This algorithm was trained to specifically identify and label boats and ships of a minimum of 7 meters in length. However, smaller vessels can also sometimes be detected.
This algorithm is compatible with Maxar imagery that meets the following criteria:
Compatible sensors: WorldView-1, WorldView-2, WorldView-3, GeoEye-1
50cm/pixel panchromatic imagery
Pansharpening OFF, Dynamic Range Adjustment (DRA) OFF, Atmospheric Compensation (AComp) OFF
Note that this algorithm does not consider the following features to be ships:
Other stationary objects in water
Ships that are on land (with the exception of dry docks near water)
This model was trained on over 25,000 image chips of Maxar 50cm, orthorectified imagery across a wide variety of nations/states, geographies, biomes, seasons, and climates on all populated continents. The training data included open-water imagery, as well as imagery covering over 50 major ports across the globe.
This model was validated against CrowdAI’s internal Maxar validation set, a globalized set of imagery drawn from the WorldView and GeoEye constellations.
See how quickly you can deploy and run models, connect to pipelines, autoscale resources, and integrate into workflows with Modzy—the ModelOps and MLOps platform