Much of the imagery and remote sensing analysis in the Open Source Community pertains to North Korea’s nuclear weapons pathway and military capability. However, many questions remain regarding economic and agricultural health in a nation known for denial of access to outside observation. But by applying emerging analytical and processing technology of satellite imagery and data, we can address the challenge of examining economic and environmental patterns in the North.
Machine Learning technology has been used to analyze rudimentary objects like roads or buildings on satellite imagery for years, but has yet to be successfully employed to better understand nuanced patterns of life. In our partnership with the analytics company Orbital Insight, we have undertaken a project of counting thousands of objects in satellite images taken over the past five years to uncover North Korea’s trade relationship with China.
This project includes counting number of trucks at each side of the Sino-Korean Friendship Bridge as a measure of trade activity between North Korea and China. By applying artificial intelligence to more than 300 satellite images, we observed fluctuations of truck counts, which peak during the month of November. A significant drop in the truck counts during the year of 2020 is noticed as a result of restricted traffic from the global pandemic, although as much as 30 trucks were observed in the month of June on both sides of the border. The project demonstrates the utilities of machine learning in analyzing emerging datasets. Careful monitoring of trade between the two states can aid in better understanding the China-North Korea economic relationship and how it evolves over time.
CISAC is also partnering with international organizations and geospatial systems specialists to apply data derived from public space mapping systems to better understand macro-environmental, agricultural, and water security trends over the past twenty years in North Korea. For decade