Uav Vision Analytics

What it does

When your agent uses it

• : creating UAV/drone vision analytics stacks that detect objects from aerial video (file
• RealSense camera
• Overlay live MAVLink telemetry (GPS
• Heading) on the annotated RTSP stream

Example prompts

• “/uav-vision-analytics”

Requirements

• Docker
• Compatibility (from SKILL.md): Requires Docker + Docker Compose v2, Intel CPU (optionally GPU/NPU with video/render groups). For pymavlink mode: PX4 SITL runs in simulation. For UAVSDK mode: uav-mission-compute-sdk must be running first. Ports 8081 (REST), 8555 (RTSP), 1883 (MQTT), 14541/udp (MAVLink) must be free. Tested with intel/dlstreamer-pipeline-server:2026.1.0 image.

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

1. 1Read this file end-to-end.
2. 2Ask the questions in ONE batched message (defaults shown in brackets); accept
3. 3Validate parameters before generating files.
4. 4Load reference files on demand — do not load all up front.
5. 5Generate the application files and verify against the completion criteria.

What it can do on your machine

Read from SKILL.md and the folder at commit 6e2ba00. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Shell commands in SKILL.md call:

  • make
  • docker
  • curl
  • pytest

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md. Its commands use docker and curl, which can reach the network depending on how they are called.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

• Compatibility

  Requires Docker + Docker Compose v2, Intel CPU (optionally GPU/NPU with video/render groups). For pymavlink mode: PX4 SITL runs in simulation. For UAVSDK mode: uav-mission-compute-sdk must be running first. Ports 8081 (REST), 8555 (RTSP), 1883 (MQTT), 14541/udp (MAVLink) must be free. Tested with intel/dlstreamer-pipeline-server:2026.1.0 image.

  From compatibility in the SKILL.md frontmatter.

Context cost

Uav Vision Analytics loads about 2.6k tokens when it runs, and up to ~11k if it reads all its reference files. Until then it costs about 181 tokens; SKILL.md has 849 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~181

When it runs · the whole SKILL.md, loaded when a task matches

~2.6k

With references · SKILL.md plus every file in references/, read only if the agent opens them

~11k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

• NoteMentions a .env fileSKILL.md:128

  - Never hardcode secrets — use `.env` variables for `HOST_IP`, device GIDs, credentials.

• NoteMentions a .env fileSKILL.md:144

  exports every variable from `.env`; an empty `http_proxy=` overrides the

• NoteMentions a .env fileSKILL.md:155

  ├── .env                             # HOST_IP, image tags

• NoteMentions a .env fileSKILL.md:186

  1. `make init` succeeds: `.env` created with auto-detected `HOST_IP` and GPU/NPU device paths.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from open-edge-platform/edge-ai-suites at commit 6e2ba00, republished under its Apache-2.0 licence (© open-edge-platform). 849 words, ~2,559 tokens.

name

uav-vision-analytics

description

Build an end-to-end UAV object detection and telemetry overlay application on Intel hardware using DL Streamer Pipeline Server with MAVLink telemetry. USE FOR: creating UAV/drone vision analytics stacks that detect objects from aerial video (file, RealSense camera, or RTSP feed), overlay live MAVLink telemetry (GPS, altitude, speed, heading) on the annotated RTSP stream, and support autonomous pipeline start/stop triggered by the drone armed/disarmed state. Supports two deployment modes: pymavlink (self-contained with PX4 SITL) and UAVSDK (integrates with uav-mission-compute-sdk). DO NOT USE FOR: ground-based camera analytics without MAVLink telemetry, cloud-only deployments, or model training.

compatibility

Requires Docker + Docker Compose v2, Intel CPU (optionally GPU/NPU with video/render groups). For pymavlink mode: PX4 SITL runs in simulation. For UAVSDK mode: uav-mission-compute-sdk must be running first. Ports 8081 (REST), 8555 (RTSP), 1883 (MQTT), 14541/udp (MAVLink) must be free. Tested with intel/dlstreamer-pipeline-server:2026.1.0 image.

license

Apache-2.0

<!-- SPDX-FileCopyrightText: (C) 2026 Intel Corporation -->

<!-- SPDX-License-Identifier: Apache-2.0 -->

Video Source (file/RealSense/RTSP)
    │
    ▼
DL Streamer Pipeline Server
  ├── gvadetect (OpenVINO YOLO11s, CPU/GPU/NPU)
  ├── gvapython (telemetry overlay — altitude, speed, heading, GPS)
  ├── gvametaconvert → gvametapublish → MQTT
  └── appsink → RTSP :8555
         │
         ▼
    QGC / ffplay / browser

MAVLink/MQTT → Pipeline Manager → start/stop pipelines on ARMED/DISARMED

{{STACK_DIR}}/
├── docker-compose-pymavlink.yml     # or docker-compose-uavsdk.yml
├── 10040_sihsim_quadx.post          # PX4 airframe MAVLink routing (pymavlink only)
├── .env                             # HOST_IP, image tags
├── .env.example                     # template copied by make init
├── Makefile                         # init, model, stack up/down, pipeline start/stop
├── configs/
│   └── config-{{PIPELINE_PREFIX}}.json   # DL Streamer Pipeline Server pipeline definitions
├── gvapython/
│   └── telemetry-overlay-{{MODE}}.py     # gvapython telemetry overlay
├── scripts/
│   └── pipeline_manager.py               # armed/disarmed pipeline lifecycle
├── mavlink-router/
│   ├── Dockerfile                         # self-contained build (pymavlink only — never reference an external path)
│   └── main.conf                          # mavlink-router config (pymavlink only)
├── resources/
│   ├── models/yolo11s/          # exported OpenVINO model
│   └── videos/gazebo.avi                 # sample video (file source)
└── tests/
    ├── conftest.py
    ├── test_stack_up.py
    ├── test_pipeline_start.py
    ├── test_rtsp_stream.py
    └── test_mavlink_trigger.py

© open-edge-platform, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 10 other files (references) in federal-and-aerospace-ai-suite/uav-vision-analytics/.github/skills/uav-vision-analytics of open-edge-platform/edge-ai-suites.

• SKILL.md
• benchmark.md
• evals/evals.json
• example-prompts/01-pymavlink-sim-all-devices.md
• example-prompts/02-mavsdk-three-camera.md
• example-prompts/03-realsense-gpu.md
• references/DEPLOY.md
• references/MODEL.md
• references/PIPELINE.md
• references/TELEMETRY.md
• references/TESTS.md

Open the folder on GitHubat commit 6e2ba00