Artificial intelligence helps satellite maps improve clarity, showing global renewable energy projects and forest coverage

According to news on September 3, the Allen Institute for AI, founded by Microsoft co-founder Paul Allen, recently released a new tool called Satlas, which contains the world’s first Maps using generative artificial intelligence techniques to enhance the clarity of satellite imagery can show renewable energy projects and forest cover around the world.

This site noticed that this map uses satellite images from the European Space Agency’s Sentinel-2 satellite. However, these images still couldn't clearly show ground details, so they used a solution called "Super-Resolution." Basically, deep learning models are used to fill in details, such as what a building might look like, to produce a high-resolution image.

The above picture is a high-resolution image of Nakuru, Kenya, generated by artificial intelligence, and the below picture is a satellite shot of the same location Low-resolution image

Currently, Satlas focuses on renewable energy projects and forest cover around the world. The data is updated monthly and covers the portion of the Earth monitored by Sentinel-2, which includes most of the world except Antarctica and the high seas far from land.

The map shows solar farms and onshore and offshore wind turbines and can also be used to see changes in tree canopy cover over time, which is useful for policymakers trying to achieve climate and other environmental goals. Very important.

According to the Allen Institute, this is the first tool with such broad coverage that is free and open to the public, and its developers also said that this may be the first demonstration of super-resolution technology in a global map .

Of course, there are still some problems that need to be solved. Like other generative AI models, Satlas is prone to "hallucinations" and will sometimes draw buildings in a strange way. For example, the building is rectangular, while the model may think it is a trapezoid or something, which may It is due to differences in architectural styles in different regions that the model is difficult to predict. Another common "illusion" is placing cars and boats where the model thinks there should be, based on the images used to train the model.

To develop Satlas, the Allen Institute team had to manually browse satellite images, marking 36,000 wind turbines, 7,000 offshore platforms, 4,000 solar farms and 3,000 tree canopy coverage . For super-resolution, they fed the model many low-resolution images taken at different times from the same place. The model uses these images to predict sub-pixel details in high-resolution images.

The Allen Institute also plans to expand Satlas to provide other types of maps, including one that can identify the types of crops grown around the world.

The above is the detailed content of Artificial intelligence helps satellite maps improve clarity, showing global renewable energy projects and forest coverage. For more information, please follow other related articles on the PHP Chinese website!