Bharat in Betastartup field guide
← Back to the atlas
No. 1157Problem statementSpace and research

Development and Optimization of Al model for Feature identification/ Extraction from drone orthophotos.

Source

Source summary

Develop a cloud AI/ML model extracting buildings, roads and water bodies from SVAMITVA drone orthophotos. Classify roofs as RCC, tiled, tin or other, supporting uses such as solar or property-tax assessment. Use labelled data from ten villages for training/validation, separately test performance and optimise speed/resources. The source sets a 95% identification-accuracy target, not an achieved result. Prepare integration into the ministry’s existing systems, with architecture, training, deployment and evaluation documentation.

No. 528Problem statementSpace and research

Analysis of SAR imagery using various tech to include SAR & EO Imagery fusion and Metadata exploitation.

The source notes that synthetic-aperture-radar imagery remains useful at night and in bad weather but is hard to interpret without optical references and specialist knowledge. It suggests, rather than mandates, fusing SAR and electro-optical imagery, extracting changes, and using user-defined stable terrain, infrastructure or standard targets to enhance readability. It also asks for software that exposes material-related information derivable from SAR metadata while keeping observations distinguishable from enhanced or substituted content.

No. 592Problem statementSpace and research

Training Simulator for Space Activities

Develop a software simulator for periodic training on dynamic space-based contingency scenarios and for testing proposed countermeasures once they exist. The simulator should be scalable so additional feasible scenarios and their countermeasures can be incorporated over time; the page does not define scenario content or evaluation metrics.

No. 622Problem statementSpace and research

Development of an Indigenous Tool for Space Domain Training Surveillance and Operations in the Indian Air Force (IAF)

Develop an indigenous Indian Air Force tool for space-domain training, surveillance and operations. It must support user-defined volumetric scans, proximity screening around Indian satellites, conjunction and collision analysis, coverage-gap and manoeuvre detection, space-object correlation and cataloguing, re-entry assessment and rapid processing of selected sub-orbital activity. The software and hardware must run in the IAF network, accept open, high-definition ephemeris and authorised internal data, follow the required development and certification process, scale to later capabilities, remain maintained by the developer and use perpetual licensing. Sensitive operational details are intentionally omitted.