Show HN：一个能听鸟叫并将其绘制成 19 世纪风格插画的电子墨水屏相框

Show HN：一个能听鸟叫并将其绘制成 19 世纪风格插画的电子墨水屏相框
Show HN: An e-ink frame that hears birds and draws them as 1800s illustrations

原始链接: https://github.com/arnegiacomo/fugleramme

**Fugleramme** 是一款基于树莓派（Raspberry Pi）的数字观鸟相框。它能够实时捕捉并识别本地鸟类的鸣叫声，并将识别结果转化为精美的 19 世纪手绘自然历史插图进行展示。该软件使用 **BirdNET-Go** 来监听和识别鸟鸣，并将识别到的物种与包含 800 多幅公有领域插图的精选库进行匹配。这些图像会根据鸟类的实际体重，动态调整比例并渲染到电子墨水屏（或网页端信息亭）上。当未检测到鸟类时，相框会显示一张优雅的“空枝”图案。该系统专为 DIY 爱好者设计，完全在本地运行，确保了隐私与完全的自主控制。它目前针对斯堪的纳维亚、英国和中欧的鸟类进行了优化，并正在努力扩展对北美地区的支持。该项目具有高度的模块化：你可以使用树莓派、13.3 英寸 Inky Impression 显示屏和一个麦克风自行组装，或者直接在任何设备上以网页信息亭模式运行该软件。详细的文档、容器化安装指南和贡献准则均可在 GitHub 上找到。无论你是选择亲手打造，还是购买预装成品，Fugleramme 都能通过艺术与科学的结合，让你的花园焕发生机。

A Hacker News thread discusses "Show HN: An e-ink frame that hears birds," a project by Arne Munthe-Kaas. The device uses a microphone to detect local bird calls via the open-source tool BirdNET-Go, subsequently displaying authentic 1800s-style public domain illustrations of the identified birds on an e-ink screen. Key technical features include local operation on a Raspberry Pi, intelligent scaling based on bird mass, and dithering for e-ink optimization. While some users debated whether the project was a derivative of existing work like "AvianVisitors," the author clarified that his project functions independently as a companion to BirdNET-Go and prioritizes human-made, historical artwork over AI-generated imagery. The project has garnered significant interest for its creative application of technology, though commenters noted that large e-ink displays remain prohibitively expensive. The author continues to refine the project, with current bird species coverage focused primarily on Northern Europe.

原文

E-ink bird frame for Raspberry Pi - real-time bird detection by audio, fully local AI, rendered as real, hand-cut 1800s bird illustrations.

Sorry about the dirty window - squirrels have been stealing the bird food.

Note

Still in early development: expect the odd bug and a few unpolished edges, with plenty more features to come.

Live on fugleramme.arnegiacomo.dev running from my kitchen window and displaying the actual birds currently heard in my garden (Bergen, Norway).

Hardware, install and operations docs: arnegiacomo.dev/fugleramme

BirdNET-Go listens on a mic and handles the classifier. Fugleramme polls its api, matches each species to an illustration, then packs them onto a page, and redraws only when the birds change - on an Inky Impression e-ink panel, and as a web kiosk serving the same view. There's an admin page that lets you configure what to show, and automatic updates and such.

If you already run BirdNET-Go, point the frame at it instead - on the same machine or anywhere else reachable from your network.

Tip

The e-ink panel is not required, although it's recommended for the intended experience. Without one, Fugleramme runs web-only - show the kiosk on a display over HDMI, or open it from any device on the network.

A Raspberry Pi 5, an Inky Impression 13.3" (Spectra 6), a mic and an A4 frame. Full parts list, recommendations and alternatives: Hardware.

Half the point of this project is showing off some amazing public-domain natural-history illustrations. Over 800 cut-outs covering more than 400 species, every one taken from a real plate and hand-curated for this project (no art is AI-generated, though some has been retouched with AI).

Each detected species is matched to its illustration, background-removed, and packed onto a textured paper page with the larger birds toward the centre, sized by body mass. An empty window shows a bare perch.

The plates are Scandinavian, British and central European, so the Nordics, the British Isles and Germany are best covered. Elsewhere not so much (yet). Broader European and North American coverage is in the works!

See Adding artwork for manual cutout steps.

Run locally (for development)

uv sync                                       # set up venv
uv run fugleramme-fake-detector               # stand-in BirdNET-Go on :8090
uv run fugleramme-dev                         # start service on :8080 with hot-reload

The fake detector's flags, and working against a real station instead: Running it without a Pi.

Install on a Raspberry Pi

From the pi (assuming you have the hardware up and running):

curl -fsSL https://raw.githubusercontent.com/arnegiacomo/fugleramme/main/install.sh | bash

Asks where BirdNET-Go should live and which ports to use, clones the repo, installs the required deps, and starts the frame as a systemd service. NB! Will probably require a reboot on a fresh system.

From a blank SD card, see the full install guide.

docker run -d -p 8080:8080 -v fugleramme:/data \
  -e FUGLERAMME_DETECTOR_URL=http://birdnet.local:8080 \
  ghcr.io/arnegiacomo/fugleramme

Or build the image from a checkout:

docker build -t fugleramme .
docker run --rm -p 8080:8080 -v fugleramme:/data \
  -e FUGLERAMME_DETECTOR_URL=http://birdnet.local:8080 fugleramme

Kiosk on :8080, admin on :8080/admin, everything it persists in /data.

On a Linux box with a USB mic, this brings up BirdNET-Go alongside it:

curl -fsSL https://raw.githubusercontent.com/arnegiacomo/fugleramme/main/examples/docker-compose.yml -o docker-compose.yml
docker compose up -d

See Container for more info.

Contributions are very welcome and encouraged - fixes, docs and artwork most of all. Thanks to everyone who has contributed so far ❤️

Something is broken - a bug report
A question, an idea, or a frame you have built - the FAQ first, then Discussions
A fix, a doc change, or a bird you have cut - open a PR, no issue needed

See Contributing for more info.

Code: MIT - see LICENSE.
Detection (BirdNET-Go, installed separately as a container): CC BY-NC-SA 4.0, non-commercial only. BirdNET model by the Cornell Lab of Ornithology and Chemnitz University of Technology, taxonomy data powered by eBird.org.
Bird images: each style folder carries its own terms and sources, and its manifest links the plate every file was cut from. classic is CC BY-SA 4.0 - see assets/artwork/classic/ATTRIBUTION.md.
Label fonts (assets/fonts/): SIL OFL 1.1 - see assets/fonts/ATTRIBUTION.md.
Bird sizes (assets/bird_sizes.csv): body mass from AVONET (Tobias et al. 2022, Ecology Letters, doi:10.1111/ele.13898), CC BY 4.0.
BirdNET scientific-name aliases (assets/birdnet_aliases.json): OpenFauna's compiled taxonomic alias map, CC BY-SA 4.0 - see assets/ATTRIBUTION.md.

I've built a few of these. If you'd like one rather than building it yourself, please get in touch.