Targetless Joint Extrinsic, Temporal, and Focal-Length Calibration of Roadside Radar–Camera Systems from Traffic

Roadside millimetre-wave radar–camera units need an extrinsic calibration that is difficult to obtain and maintain above live lanes. This paper presents a targetless calibration framework whose only inputs are ordinary traffic, a coarse mount constraint and a nominal camera model. Correspondences are decided by a consensus over road-plane homography hypotheses, the radar’s range scale is checked against its Doppler channel, and the rigid transform, time offset and focal length are estimated jointly, with one focal length per preset and the lever arm confined to the mount box. On seven recordings from one roadside unit, a prospective vehicle-grouped protocol gives 5.1 to 7.3 pixels of reprojection error on the 86 to 95% of test pairs its gate retains and 6.1 to 9.5 pixels on all; PnP with the native focal length is within 0.4 pixels. The focal lengths of the three presets agree within 1% and lie 2 to 3.5% below the native value and an aerial-imagery reference, the time offsets agree within 0.022 s, calibrations transfer between recordings, and the rotation agrees with that reference to 1.7°. Planar traffic constrains the lever arm along the optical axis only weakly; the mount bound confines it without establishing the true translation.

Authors

Institutions

Publication Details

Journal
Sensors
Published
2026-11-12
DOI
https://doi.org/10.3390/s26227150
Primary Topic
Autonomous Vehicle Technology and Safety
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Targetless Joint Extrinsic, Temporal, and Focal-Length Calibration of Roadside Radar–Camera Systems from Traffic

Zeynab Rokhi, Ali Emadi
Sensors
Autonomous Vehicle Technology and Safety
article

Targetless Joint Extrinsic, Temporal, and Focal-Length Calibration of Roadside Radar–Camera Systems from Traffic

Zeynab Rokhi, Ali Emadi
article en

Abstract

Roadside millimetre-wave radar–camera units need an extrinsic calibration that is difficult to obtain and maintain above live lanes. This paper presents a targetless calibration framework whose only inputs are ordinary traffic, a coarse mount constraint and a nominal camera model. Correspondences are decided by a consensus over road-plane homography hypotheses, the radar’s range scale is checked against its Doppler channel, and the rigid transform, time offset and focal length are estimated jointly, with one focal length per preset and the lever arm confined to the mount box. On seven recordings from one roadside unit, a prospective vehicle-grouped protocol gives 5.1 to 7.3 pixels of reprojection error on the 86 to 95% of test pairs its gate retains and 6.1 to 9.5 pixels on all; PnP with the native focal length is within 0.4 pixels. The focal lengths of the three presets agree within 1% and lie 2 to 3.5% below the native value and an aerial-imagery reference, the time offsets agree within 0.022 s, calibrations transfer between recordings, and the rotation agrees with that reference to 1.7°. Planar traffic constrains the lever arm along the optical axis only weakly; the mount bound confines it without establishing the true translation.

SensorsVol. 26(22)
McMaster University (CA)
Autonomous Vehicle Technology and Safety
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon