Satellite- and AI-powered detection of artisanal and small-scale mining sites


Artisanal and small-scale mining (ASM) is responsible for 10% of global gold production, with an estimated market value of US $14bn, affecting the lives of 10 to 20 million miners (and an additional 80 million miners for other raw materials) and their families. ASM sites are located in more than 80 developing countries and associated with lower environmental and work safety standards. Certification of compliant ASM operations, enforcing concession rights, sustainability and ecological standards play an essential role to create wealth for local communities and protecting the environment and workers. However, as most ASM operations are dispersed in large, remote areas such as the Amazon, control of ASM sites is quite challenging, resulting in a large number of illegal and informal mines that are not compliant to national and international work safety and environmental standards.

AIASMSpotter wants to develop a software that enables different users/customers and stakeholders to identify and monitor ASM gold sites automatically from satellite imagery such as Sentinel- 1 (radar) and Sentinel-2 (optical) by applying Deep Learning.


The first main task is the acquisition of an adequate training database. The database consists of examples of input-output pairs of the model. Personnel will locate and identify ASM activities in input data and label them with appropriate software tools. After the data is gathered, the Deep Learning phase begins. Literature research is conducted on the respective algorithms that solve similar tasks. Once a suitable algorithm is identified, the model will be adapted to the requirements of ASM sites. In the last phase, software development aspects are sorted out, i.e. the means and format of shipping the prototype to the user. This includes determining the software dependencies and the mode of user interaction with the model, e.g. via a jupyter notebook, a command-line tool or a software library.


AIASMSpotter is a tool for national and international authorities and organisations active in controlling ASM activity and adherence to concession rights. Moreover, AIASMSpotter is a tool for surveillance, allowing these organisations to spot illegal activities at an early stage, reaping more wealth out of ASM activity for authorities and the local population. Mining companies might use the tool to generate insights into ASM and possible exploration sites for sustainable mining.

Key Data

  • Level: International
  • Role: Subcontractor
  • Partners: 1
  • Duration: 6 months

Research Areas


dida Datenschmiede GmbH

Funded by: