A series of pre-build surveys: structured passes over a research and engineering landscape before committing to a build. Each post inventories what is currently shipping, what is publishable but not deployable, what is standardized in the surrounding ecosystem, and what fits on the hardware I actually have. The goal is to give myself, and anyone considering similar work, a starting bibliography and a sense of where the field is heading over the next twelve months.
Posts in this series
- Drone-to-Satellite Visual Geo-Localization: State of the Art, Open Source, and Edge Inference in 2026 - A May 2026 survey of cross-view localization for UAVs: NVIDIA Isaac ROS, the CVPR 2026 VGGT cluster, MAVLink integration, and what fits on Jetson Orin.
- Visual-Inertial Odometry for Drones: State of the Art, Foundation-3D, and the cuVSLAM Squeeze in 2026 - A May 2026 survey of VIO: where production stacks have consolidated, the foundation-3D + IMU research cluster, the standardized MAVLink wire format, and why building another open VIO library in 2026 is unlikely to pay off.