Planet’s Data Leveraged to Identify Dark Vessels and Monitor the Illicit Russian Oil Trade

With advances in artificial intelligence and analytics, we are just beginning to see the full potential of how Planet’s robust datasets can be used to track dark vessels globally and uncover illicit supply chain routes. Overlaying machine learning algorithms on our frequently captured satellite data enables government bodies, coast guards, border security, and intelligence agencies to monitor ports and coastlines, detect and track vessels, and uncover suspicious activity in contested waters. 

Ships have a registered ship identity, a digital identity (electronic signature), and a physical identity; and analysts have to look at all three to investigate them and understand vessel spoofing operations, where ships disguise themselves as legitimate by broadcasting inaccurate information. This is why satellite imagery can be so essential. While ships may be able to acquire a new registered identity, or broadcast a different digital identity on the automated identification system (AIS), it is harder to make significant structural changes that fundamentally alter a ship’s physical identity. Planet’s global satellite imagery can reveal these ship details as well as their near-daily location, offering critical information for investigating illicit ship-to-ship transfers and port stops. 

Planet’s automated Vessel Detection Feed combines our PlanetScope data with data analysis tools specifically designed to detect maritime vessels and track ship movement in strategic areas of interest. The near-daily imagery helps fill in intelligence gaps to monitor vessels even in the absence of AIS data. Beyond our own analytical feeds, our users are using their own AI algorithms to make critical discoveries on dark vessels. C4ADS and RUSI used Planet’s satellite data to conduct investigative studies on North Korea’s illicit oil procurement networks, and AI company, Windward, also combined their AI vessel tracking data with our imagery to report on the alleged laundering of Ukrainian grain by Russian dark vessels. Our images have also enabled journalists to investigate ship-to-ship transfers, allegedly allowing oil to flow to Russia despite sanctions.

Recently, our partner Royce Geo leveraged our imagery, paired with their computer vision models, to uncover critical answers to how tanker traffic volumes from Russia are changing, and where they are exporting and importing unsanctioned Russian oil across the globe. Using their models, Royce Geo extracted ship data and classified tankers. They then fused AIS data with satellite imagery to locate vessels that were potentially subverting their activity. Their analysis found that more tankers coming from Russian ports are now heading towards Chinese export destinations. This analysis demonstrated how applying computer vision models with PlanetScope data enabled the research team to quickly detect Russian tankers operating with their AIS transmitted disabled (referred to as operating “dark”) as they allegedly traveled between Russia and China to hit ports and make ship-to-ship transfers. 

To learn about their research process and findings in full, read their article, cross-posted below:

###

Data, Analytics, and Tankers Full of Russian Oil

Global interest continues to grow in understanding how Russia continues to export oil and natural gas products while most nations have imposed sanctions and price caps. This issue is deeply impactful to the global economy, the continued conflict in Ukraine, and concerns about a growing risk of potential accidents and ecological impacts.  

Our team here at RoyceGeo is collaborating with Spire and Planet to show how analytics applied to remotely sensed Automated Information System (AIS) vessel broadcasts and satellite imagery can answer key questions about magnitudes, trends, and specific vessel and fleet activities. While we have by no means conducted an exhaustive study of such a large and complex problem, we have been able to quickly show that remotely sensed data paired with analytics can answer critical questions about how tanker traffic volumes from Russia are changing, which ports are exporting and importing Russian oil, the feasibility of using remotely sensed data to find tankers that are subverting their activity, and the prevalence of older vessels that have recently started operating from Russian ports.  

To illustrate how we approach problem sets with data and analytics, we connected Royce Geo CURVE Operational Environment (CURVE OE) with Spire and Planet API’s to rapidly ingest AIS and imagery data for the relevant time-period and geographic regions. We applied a Computer Vision (CV) model to extract ships and classify tankers within satellite images. Then we fused the AIS data with imagery-based tanker detections to locate vessels potentially subverting their activity by not transmitting or altering their AIS broadcast.  

Here are some of the interesting things we found:

1. Analytics applied to global AIS data collections show that since sanctions were imposed, vessel traffic has decreased at some large Russian ports with petroleum / oil / lubricant (POL) export capabilities, while some smaller Russian ports specializing in liquid natural gas (LNG) exports have shown increasing tanker traffic.

2. AIS data, which is collected from vessel transmissions, shows that the global fleet of large LNG and POL tankers is so large that detecting a significant shift of vessel operations to Russian ports is challenging (bottom left chart). The shift is subtle and is also happening at least in part with dark and gray vessels that disable their AIS transmissions some or all the time. Analytics applied to global AIS data do show that tankers originating from Russian ports have shifted the destination for exports, mainly to China (top left chart).

3. Analysis of AIS and satellite imagery shows a recent and growing buildup of vessels anchored in the Eastern Mediterranean, near the entrance to the Black Sea.  

4. Computer Vision technology applied to PlanetScope daily satellite imagery collections quickly detected LNG and POL tankers operating with their AIS transmitters disabled. The two examples below show an LNG and a POL vessel that largely operated “dark” following stops in Russia and were imaged while making port calls in China and performing ship-to-ship transfers.  

5. Analysis of AIS data does show a probable high-risk trend of older tankers have returned to service and started exporting Russian oil across the Baltic Sea.  

6. Analysis of AIS and Satellite imagery provides evidence of specific vessels transitioning ownership and shifting operations to Russian ports from other parts of the globe.  

These are just a few of the facets of such a geographically massive and yet subtle problem that we have studied. To learn more about Royce Geo and our unique capabilities to deliver decision ready information from data, check us out here: https://www.roycegeo.com/