INPE approves R$118 million project for satellite, land and environmental AI tools

The National Institute for Space Research (INPE) said on October 23 that its IdeIA Terra Brasil-INPE project had been approved under the federal IdeIA Terra Brasil program. The four-year initiative has R$118,063,302.44 in planned resources and aims to develop artificial-intelligence tools for satellite-data processing, environmental monitoring and analysis of land records. This is an approval announcement, not evidence that tools already operate nationally or have produced measured changes in deforestation or land registration.

The first of three subprojects would expand Brazil Data Cube, an INPE-developed platform for processing and distributing remote-sensing data. The planned addition is AI capability for those processing and distribution functions. The release specifies no target number of satellite scenes, processing speed, spatial resolution or accuracy improvement, so those performance measures cannot be inferred from the project description.

The second subproject, BiomasBR-IA, is intended to incorporate machine-learning techniques into mapping of deforestation and land use and land cover. The release does not provide a new deforestation rate, revise a historical series or forecast future land change. Planned mapping tools are not themselves evidence that deforestation has risen or fallen; results would need to be assessed after the work produces published data.

The third effort, IA-CAR, is designed to automate stages of evaluation and prioritize records in Brazil’s Cadastro Ambiental Rural, the Rural Environmental Registry. INPE describes this as support for reviewing registry information. It does not say automation will replace legal review, determine ownership, certify environmental compliance or regularize land claims. The announcement gives neither a count of records to process nor an expected reduction in review time.

The project is part of the Brazilian Artificial Intelligence Program’s public-service-improvement axis and results from a technological procurement commissioned by the Ministry of Science, Technology and Innovation, with funding from the National Fund for Scientific and Technological Development. The research, development and innovation agreement was signed by the National Education and Research Network (RNP), INPE and FUNCATE. RNP coordinates the wider program; INPE says its effort will integrate Earth observation, geoinformatics, AI and data and supercomputing infrastructure teams.

The three workstreams address separate layers of an information chain: satellite-data infrastructure, machine-assisted mapping and prioritization of rural environmental records. INPE says the intended public value is to make environmental, territorial and land information more available to public managers. The announcement gives total planned funding and four-year duration, but no detailed milestone schedule, procurement breakdown or performance baseline. Outcomes should be measured later against such criteria, not treated as accomplished at approval.