vision-calibration

High-level entry crate and facade for the calibration-rs toolbox.

This is the recommended crate for most users. It re-exports all sub-crates through a unified API.

Features

Session API: Structured calibration workflows with step functions, state tracking, and JSON checkpointing
6 workflows: Planar intrinsics, Scheimpflug intrinsics, single-camera hand-eye, rig extrinsics, rig hand-eye, laserline device
Prelude module: Minimal imports for planar "hello world" calibration
Foundation access: Direct access to core types, linear solvers, and optimization when needed

Quick Start

Add to your Cargo.toml:

[dependencies]
vision-calibration = { git = "https://github.com/VitalyVorobyev/calibration-rs" }

Planar Intrinsics Calibration

use vision_calibration::prelude::*;
use vision_calibration::planar_intrinsics::{step_init, step_optimize};

let mut session = CalibrationSession::<PlanarIntrinsicsProblem>::new();
# let dataset: PlanarDataset = unimplemented!();
session.set_input(dataset)?;

step_init(&mut session, None)?;
step_optimize(&mut session, None)?;

let result = session.export()?;

Single-Camera Hand-Eye Calibration

use vision_calibration::session::CalibrationSession;
use vision_calibration::single_cam_handeye::{
    SingleCamHandeyeProblem,
    step_intrinsics_init, step_intrinsics_optimize,
    step_handeye_init, step_handeye_optimize,
};

let mut session = CalibrationSession::<SingleCamHandeyeProblem>::new();
# let input = unimplemented!();
session.set_input(input)?;

step_intrinsics_init(&mut session, None)?;
step_intrinsics_optimize(&mut session, None)?;
step_handeye_init(&mut session, None)?;
step_handeye_optimize(&mut session, None)?;

let result = session.export()?;

Scheimpflug Intrinsics Calibration

use vision_calibration::core::PlanarDataset;
use vision_calibration::session::CalibrationSession;
use vision_calibration::scheimpflug_intrinsics::{
    ScheimpflugIntrinsicsConfig, ScheimpflugIntrinsicsProblem, run_calibration,
};

# fn main() -> anyhow::Result<()> {
# let dataset: PlanarDataset = unimplemented!();
let mut session = CalibrationSession::<ScheimpflugIntrinsicsProblem>::new();
session.set_input(dataset)?;

let config = ScheimpflugIntrinsicsConfig::default();
run_calibration(&mut session, Some(config))?;
let result = session.export()?;
println!("mean reprojection error: {:.4}", result.mean_reproj_error);
# Ok(())
# }

Using the Prelude

use vision_calibration::prelude::*;

// Minimal hello-world imports:
// CalibrationSession + planar problem + planar runner + core planar types.

Available Problem Types

Problem Type	Steps
`PlanarIntrinsicsProblem`	`step_init` → `step_optimize`
`SingleCamHandeyeProblem`	`step_intrinsics_init` → `step_intrinsics_optimize` → `step_handeye_init` → `step_handeye_optimize`
`RigExtrinsicsProblem`	`step_intrinsics_init_all` → `step_intrinsics_optimize_all` → `step_rig_init` → `step_rig_optimize`
`RigHandeyeProblem`	All 6 steps (intrinsics + rig + hand-eye)
`LaserlineDeviceProblem`	`step_init` → `step_optimize`
`ScheimpflugIntrinsicsProblem`	`step_init` → `step_optimize`

Each problem type also provides a run_calibration convenience function that runs all steps.

Module Organization

Module	Description
`session`	Calibration session framework (`CalibrationSession`, `ProblemType`)
`planar_intrinsics`	Single-camera intrinsics (Zhang's method)
`single_cam_handeye`	Single camera + hand-eye calibration
`rig_extrinsics`	Multi-camera rig extrinsics
`rig_handeye`	Multi-camera rig + hand-eye
`laserline_device`	Camera + laser plane device
`scheimpflug_intrinsics`	Single-camera planar intrinsics with Scheimpflug tilt
`core`	Math types, camera models, RANSAC
`linear`	Closed-form initialization algorithms
`optim`	Non-linear optimization
`synthetic`	Deterministic synthetic data generation
`prelude`	Convenient re-exports

When to Use This Crate vs. Sub-Crates

Use Case	Recommended
General calibration tasks	`vision-calibration` (this crate)
Only need math types/camera models	`vision-calibration-core`
Only need linear initialization	`vision-calibration-linear`
Building custom optimization	`vision-calibration-optim`
Need pipeline + JSON I/O	`vision-calibration-pipeline`

Examples

See examples/ directory:

Example	Problem Type	Data
`planar_synthetic`	Planar intrinsics	Synthetic
`planar_real`	Planar intrinsics	Real stereo images
`stereo_session`	Rig extrinsics	Real stereo images
`stereo_charuco_session`	Rig extrinsics	Real stereo ChArUco images
`handeye_synthetic`	Single-camera hand-eye	Synthetic
`handeye_session`	Single-camera hand-eye	KUKA robot data
`rig_handeye_synthetic`	Rig hand-eye	Synthetic
`laserline_device_session`	Laserline device	Session API demo

vision-calibration 0.3.0