ardupilot-binlog

Parser for ArduPilot DataFlash BIN log files.

Reads .bin / .BIN files produced by ArduPilot's onboard DataFlash logger. The format is self-describing — message schemas are discovered from FMT messages within each file, so this crate works with any ArduPilot version without hardcoded message definitions.

Usage

[dependencies]
ardupilot-binlog = "0.2"

Parse all entries

use ardupilot_binlog::File;

let file = File::open("flight.bin")?;
let mut reader = file.entries()?;

for result in &mut reader {
    let entry = result?;
    if entry.name == "ATT" {
        let roll = entry.get_f64("Roll").unwrap_or(0.0);
        let pitch = entry.get_f64("Pitch").unwrap_or(0.0);
        println!("ATT: roll={roll}, pitch={pitch}");
    }
}

Collect into a Vec

use ardupilot_binlog::File;

let file = File::open("flight.bin")?;
let entries: Vec<_> = file.entries()?.collect::<Result<Vec<_>, _>>()?;

let gps_entries: Vec<_> = entries.iter()
    .filter(|e| e.name == "GPS")
    .collect();

for entry in gps_entries {
    let lat = entry.get_f64("Lat").unwrap_or(0.0) / 1e7;
    let lng = entry.get_f64("Lng").unwrap_or(0.0) / 1e7;
    println!("GPS: {lat}, {lng}");
}

Get time range

use ardupilot_binlog::File;

let file = File::open("flight.bin")?;
if let Some((first, last)) = file.time_range()? {
    let duration_secs = (last - first) as f64 / 1_000_000.0;
    println!("Flight duration: {duration_secs:.1}s");
}

Read from any Read source

use ardupilot_binlog::Reader;
use std::io::Cursor;

let data: Vec<u8> = /* BIN data from network, embedded resource, etc. */
# vec![];
let mut reader = Reader::new(Cursor::new(data));

for result in &mut reader {
    let entry = result?;
    println!("{}: {} fields", entry.name, entry.len());
}
# Ok::<(), ardupilot_binlog::BinlogError>(())

Inspect discovered formats

use ardupilot_binlog::File;

let file = File::open("flight.bin")?;
let mut reader = file.entries()?;
for result in reader.by_ref() { result?; }

for (_, fmt) in reader.formats() {
    println!("{}: format='{}', labels={:?}", fmt.name, fmt.format, fmt.labels);
}

Design

• Synchronous API — file parsing is CPU-bound sequential work; consumers call from spawn_blocking() if needed
• Minimal dependencies — only thiserror
• Error recovery — on corrupted data, scans forward for the next valid message header
• No hardcoded message types — all schemas discovered from FMT messages at runtime

License

MIT OR Apache-2.0