Deepen AI launches self-serve targetless calibration for lidar-camera rigs

Deepen AI on Oct. 6 launched Deepen Calibrate, a self-serve web app and API that lets autonomous vehicle and robotics teams calibrate lidar-camera systems from existing recordings without targets or a calibration bay. The product offers a free preview before payment, with launch pricing at $150 per download through Dec. 31, 2026.

Why it matters: - Calibration errors can quietly degrade perception, training labels and object tracking across autonomous systems. - Deepen AI is trying to cut the time and friction teams spend on recalibrating rigs after sensor swaps, remounts or drift. - The launch gives teams a way to review results before paying, which lowers the risk of trying a new calibration workflow.

What happened: - Deepen AI on Oct. 6 announced Deepen Calibrate, a self-serve, targetless multi-sensor calibration service for autonomous vehicles and robots. - The service is available as a web app and an API at targetless.deepen.ai. - Engineers upload an existing ROS1 .bag, ROS2 .db3 or .mcap recording, then review a free before-and-after preview, typically within 15 to 25 minutes. - Teams pay only if they choose to download the extrinsics and full quality report.

The details: - Deepen Calibrate removes the need for checkerboard, ChArUco or AprilGrid target sessions. - The service can work from a short drive through a mostly static scene, as little as 30 seconds long. - The workflow uses a rough position for each camera from CAD, a URDF or the recording's own transforms. - The system solves every sensor in the recording jointly in one run, rather than one sensor pair at a time. - The service supports recordings up to 50 GB. - The platform accepts ROS1 .bag, ROS2 .db3 and .mcap files unmodified, plus zipped ROS 2 bag folders. - Supported lidar vendors include Velodyne, Ouster, Hesai and RoboSense, along with other point-cloud lidars. - Every job returns per-sensor confidence scores, quality flags, error statistics and a visual overlay report. - Each run includes a free preview of the lidar on each camera with per-pair confidence. - Launch pricing is $150 to download the result, listed as 90% off the $1,500 list price through Dec. 31, 2026, with no monthly commitment. - Sign-in requires a work or university email. - The open-source Data Checker CLI, deepen-bag-check under the Apache 2.0 license, checks a recording locally without upload or account creation. - The service currently covers lidar-camera extrinsics only. - Multi-lidar and full-stack calibration, including IMU, GNSS/INS, radar, camera intrinsics and inter-sensor time offsets, are available through Deepen enterprise engagements. - A customer-hosted deployment is available for data-residency or export-control requirements, and an AWS Marketplace listing is coming soon.

Between the lines: - Deepen AI is positioning calibration as a product teams can self-serve instead of a specialist project. - The company is leaning on transparency tools like confidence scores, quality flags and overlays to make results easier to review. - The free preview and pay-on-download model lower adoption friction, while the enterprise path keeps higher-complexity work inside paid engagements. - Deepen AI said the targetless method is patented and based on the same engine used in its enterprise calibration engagements. - The company has calibrated production sensor rigs for automotive and robotics teams since 2017. - Mohammad Musa, Deepen AI co-founder and CEO, said teams should not need a target board and calibration bay to get calibration right and that engineers can validate a recording from a previous drive in minutes. - Musa said Deepen AI calibration engineers will work alongside any team that wants to try the product on its own data through the end of the year.

What's next: - Deepen AI said support from its calibration engineers will continue through the end of the year for teams testing the product. - The company plans to add an AWS Marketplace listing soon for customer-hosted deployments. - Deepen AI says it will continue offering broader calibration capabilities through enterprise engagements while the self-serve product focuses on lidar-camera extrinsics.

The bottom line: - Deepen AI is betting that faster, targetless calibration with a free preview will make a usually specialized workflow routine for autonomy teams.