Add a copper-rs human pose reference

copper-rs-pose/.gitignore

@@ -0,0 +1,5 @@
+.avocado/
+.avocado-state
+.DS_Store
+app/target/
+app/.cargo/

copper-rs-pose/README.md

@@ -0,0 +1,36 @@
+---
+language: Rust
+targets:
+  - imx8mp-evk
+  - imx91-frdm
+  - imx93-evk
+  - imx93-frdm
+  - jetson-agx-orin
+  - jetson-agx-thor
+  - jetson-orin-nano
+  - raspberrypi4
+  - raspberrypi5
+  - rb3gen2
+  - rubikpi3
+topics:
+  - robotics
+  - vision
+  - ai
+  - camera
+  - gstreamer
+  - pose-estimation
+  - cross-compilation
+icon: icon.png
+---
+
+# <img src="icon.png" width="32" height="32" style="vertical-align: middle;" /> copper-rs Human Pose
+
+A [copper-rs](https://github.com/copper-project/copper-rs) task graph in Rust that reads a USB camera, finds human poses with YOLOv8n-pose on the CPU, and serves the results to a [Rerun](https://rerun.io) viewer on your laptop. It is based on the upstream `cu_human_pose` example.
+
+- Cross-compile a copper-rs app (a task graph in `copperconfig.ron`, made into code at build time) into one binary
+- Camera input through GStreamer, inference with Candle, no Python and no GPU
+- Watch the live camera image and the skeletons from your laptop with `rerun --connect`
+- Write the copper-rs log to `/var/lib` on a read-only root filesystem
+- See the live task graph in the copper-rs TUI over SSH
+
+The YOLOv8 pose weights are from Ultralytics and use the AGPL-3.0 license. The build downloads them, and they are not in this repository. For a product, get a license from Ultralytics or use your own model.