Stereo Warehouse Sorting and Labeling Data

It is first-person video of a person doing sort-and-label work in a warehouse: printing labels and applying them to outbound boxes, sorting those boxes by carrier or route, sorting and separating boxes and products by type with a handheld label gun and sorting carts, and packing orders, taping them shut and applying the shipping label. It is recorded from their own point of view with two calibrated cameras, with hand pose and sensor streams alongside. The tasks are the ones the site’s catalog lists in three rows. Stereo describes the capture, and MCAP describes the container. They are separate choices.

The site does not state that stereo is already running on these rows. The catalog lists a head-mounted phone or a multiview rig as the capture setup, and a stereo pair is a capture choice agreed in the skill spec and confirmed on a Spec Pilot. The data is human capture by paid, consented contributors, not robot or teleoperation data. The site does not state that the video carries decoded barcode values, read label text, sort-correctness or accuracy labels, or a sort rate, so this page promises none. Firsthand is not affiliated with Foxglove or the MCAP maintainers.

Three, all in the Warehouse & Distribution Center environment of the Logistics & Warehousing sector, all rated Easy, and all listed as Ready to deploy. "Sort and label outbound boxes using a label printer and box cutter" reads "Print labels, apply them to outbound boxes and sort the boxes by carrier or route." "Sort, separate, and label boxes and products by type using a handheld label gun and sorting carts." is described as a "Warehouse with racked storage, sorting stations, and dedicated labeling benches." "Packing & Labeling" reads "Pack orders into boxes with filler, tape them shut and apply the shipping label."

All three list a head-mounted phone or a multiview rig as the capture setup. The catalog lists the environment examples as an e-commerce fulfillment hub, a distribution center, a cold storage facility and a freight warehouse, and does not say which one a row was captured in. None of the three is listed as Active collection, and the site says nothing further about what Ready to deploy means in volume, hours or delivered stereo footage, so this page claims none. It names only the tools and items that appear in the rows: a label printer, a box cutter, a handheld label gun, sorting carts, boxes, products, filler, tape and the shipping label. The first table below reproduces the three rows as the catalog states them.

Three nearby rows sit in a different environment and are not relied on here: "Scan + sort parcels", "Apply shipping labels" and "Cross-dock re-label" in the Warehouse Zone set are listed as Available on request.

The egocentric sorting task page covers sorting across many sectors, from returns bins to linen, and lists the two sort-and-label warehouse rows among them, with no stereo or MCAP treatment. The egocentric barcode scanning and labeling page covers scan, print, apply and tag rows across settings, including all three rows here, as egocentric catalog treatment without an MCAP treatment. The stereo order packing and taping page gives stereo and MCAP treatment to four packing and freight rows, one of which, "Packing & Labeling", also appears here.

This page is the stereo and MCAP treatment of the two sort-and-label rows together with "Packing & Labeling" as the step that applies the shipping label, with its own stream map and failure modes. The overlap with the order packing page is that one row, and this page treats it only as the labeling step. The comparison table below sets the pages side by side.

Sort-and-label work is about where a box is, where a label goes and which destination the box is headed for: a label applied to a carton face, a box placed on a sorting cart, a product separated by type at a station. A stereo pair recovers depth from the calibrated frame pair rather than from a standalone depth sensor, so the box, the label and the hands share a frame. The site gives no depth accuracy figure for these rows and no destination, sort-correctness or label-placement label, so what a buyer does with that depth is their pipeline’s question.

A stereo pair adds depth, not legibility. The site states 1080p per eye for dataset 06 and gives no figure for the smallest label text or print that stays readable, and it does not state that label text is transcribed, so small label text and print legibility are Not stated and a spec should check them on a Spec Pilot. The site also states that depth degrades on wet and reflective surfaces and ships per-frame confidence maps so a buyer can mask rather than guess. Capture at your own facility is listed as not offered.

The site states 3D hand pose of 21 joints per hand with a per-joint visibility flag, and contact and grasp events that mark when fingers meet a tool or an item. In these rows that is the hands on a handheld label gun, a label printer or a box cutter, the hands that take a box or product off a station, and the hands that place a box on a sorting cart or apply a label.

The site does not say whether the label gun, the label printer, the box cutter, a box or a label counts as the object for contact events, so settle it in the spec. Wrist cameras give a closer view of the hands, and a multiview rig adds synchronized third-person angles when the spec asks for them.

The site describes two stereo setups and does not say which one a sorting and labeling order uses. The hard-environments stack is a global-shutter 4K/60 head camera with a 120° lens, two 1080p/60 wrist cameras, a depth sensor at 848×480 and 30 fps, and 200 Hz 6-DoF IMUs. The off-the-shelf stereo head rig, dataset 06, is dual fisheye cameras at 1080p per eye, a 62 mm baseline, per-unit calibration, a 9-axis IMU and hand-tracking output.

Which rig, resolution and streams a sorting and labeling order uses is Not stated on the site, and the third table below sets what is stated beside what is not. Intrinsics and extrinsics are re-solved at the start and end of every session, and streams are hardware-triggered onto one clock with a 2.0 ms cross-stream ceiling. Bench layout, lighting, label stock, printer placement and camera clearance are Not stated, so any requirement is settled in the spec.

MCAP is an open source container file format for multimodal log data, supporting multiple channels of timestamped pre-serialized data. It is serialization-agnostic, so messages can be Protobuf, DDS (CDR), ROS, JSON and more. Its design is row-oriented and append-only, it stores schemas alongside the data, and it offers an optional index plus optional LZ4 or Zstandard chunk compression.

Reader and writer libraries exist for C++, Go, Python, Rust, Swift and TypeScript, and the mcap CLI can inspect, merge and split files. The foxglove/mcap repository is MIT-licensed; that is the MCAP project’s licence, not the licence of a Firsthand delivery. The MCAP hub page and the glossary entry cover the format in full.

Firsthand offers MCAP delivery, and dataset 06 is stated as shipping as MCAP. For these rows the site states hand pose, a stereo-derived depth stream with confidence maps, IMU, contact events, action segments and per-frame instance masks elsewhere, and states instance segmentation for warehouse collections. What it does not state is which of those are channels in an MCAP file, how they are named, what encoding they use or at what rate.

No topic name, message schema or rate is stated on this page, and none should be assumed. Name the streams that matter in the skill spec and confirm them on a Spec Pilot batch. The fourth table below sets each stream beside what the site states about it.

Box and label occlusion is the first risk. A carton, a label sheet or a handheld label gun can hide the fingers that hold it. The visibility flag on each joint is the answer the site gives, contact events mark when a hand takes hold of an item, and a spec can ask for wrist views or a second angle so a grip the head camera loses is still recorded. The site notes that reach and grasp frames are where blur and occlusion show up. An episode is rejected when the manipulated object is out of frame for more than 15 percent of it.

Small label text and print legibility is the second. Printed labels carry small text, and whether it stays readable in a frame depends on distance, angle and print quality. The site gives no legibility figure and does not state that label text is read or transcribed, so this page promises none. Fast repetitive hand motion is the third: sort-and-label work repeats the same reach, place and press many times in a row. The head camera is global-shutter, which avoids rolling-shutter skew, but the site gives no figure for blur at sorting speed, so how a stereo pair behaves at that pace is Not stated.

Mixed warehouse lighting is the fourth. The catalog describes racked storage, sorting stations and labeling benches, and the site ships transitions and hard cases uncut, including walking between rooms, exposure hunting in dim light and occluded grasps, rather than curating footage down to the clean segments. It states no handling specific to bench or aisle lighting, so how a stereo pair copes with a bright bench beside a dim aisle is Not stated. Depth is flagged rather than interpolated where it is unreliable: an episode is rejected when more than 35 percent of depth pixels in the manipulation region are invalid.

Sequence and calibration close the list. The rows chain printing a label, applying it and sorting the box, so a spec naming each step and where it starts and ends gives a sequence that can be sliced by step, while a spec saying "sort and label boxes" gives one long action segment. The rows do not say what one episode covers, so the spec decides. Stereo depth is only as good as the rectification behind it: calibration is measured at the start and end of every session under a 0.28 px RMS gate. An episode is also rejected when an undocumented bystander appears, and a reject log ships with every batch. Sort rates, throughput, carrier, facility or customer names and scanner, printer or label-gun brands are Not stated and are not claimed here.

For every delivered episode the site states that faces, plates and screens are blurred irreversibly, and that footage that cannot be de-identified is discarded. The anonymization record ships with the data, and an undocumented bystander is a stated reject criterion.

Text printed on shipping labels, such as names, addresses or tracking codes, is not listed as a blur category on the site, so this page does not claim it is blurred. The site does not state a rule specific to a labeling bench or a carrier, or how handling differs between a fulfillment hub, a cold storage facility and a freight warehouse, so this page makes no such claim. Settle it in the spec.

The October 2026 off-the-shelf catalog lists ~324,800 hours across 18 datasets, of which stereo is ~64,800 hours across 9 datasets. The site lists no off-the-shelf sorting or labeling dataset; its Logistics tag reads "Packing and packaging, plus factory floor and site work". Stereo sorting and labeling in MCAP is therefore custom collection: custom collection at scale through a network of thousands of capture providers, for a consistent, recurring data flow to labs. The off-the-shelf stereo and hand-tracking video page shows what the rig delivers, and the custom order applies it to your task.

Where coverage is deep, the median time from signed spec to first validated batch is a median of 48 hours; where it is thin, turnaround runs three to four weeks. All three rows are Ready to deploy, and the site does not say which speed applies to a stereo sorting and labeling order, so the timeline is quoted against your spec. Collections are billed per validated hour, and rejected episodes and recaptures are never billed.

Describe the steps and where each starts and ends: printing a label, applying it to an outbound box, sorting the box by carrier or route, and separating and labeling boxes and products by type with a label gun and sorting carts. Name the tools you expect to see, whether an episode covers one box, one cart or a whole shift, what counts as a reject and the variation your pipeline is missing. Then name the streams you want in the file, the topic naming and message encoding your reader expects, whether calibration ships inside the file or beside it, and whether MCAP is the only delivery or sits beside RLDS, LeRobot or WebDataset.

Most programs begin with a Spec Pilot, a small first batch that proves the spec and accept criteria on real footage before you scale. The fifth table below shows what each spec element decides.