Stereo Window and Glass Cleaning Data

It is first-person video of a person cleaning windows and glass, recorded from their own point of view with two calibrated cameras and with hand pose and sensor streams alongside. The tasks are the two rows the site’s catalog lists for it in Commercial Cleaning & Janitorial. 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. Firsthand is not affiliated with Foxglove or the MCAP maintainers.

Two, both in the Commercial Cleaning & Janitorial environment under the Facilities & Cleaning sector, and both rated Medium. "Wash and squeegee windows using extension poles and squeegees" is described as wetting windows with a pole applicator and pulling the water off with a squeegee, and is listed as Active collection. "Window & Glass Cleaning" is described as cleaning interior and exterior glass with applicators and squeegees and polishing edges, and is listed as Ready to deploy.

Both rows list a head-mounted phone or a multiview rig as the capture setup. The first table below reproduces them as the catalog states them. This page names only the tools, parts and steps that appear in those rows: extension poles, pole applicator, applicators, squeegees, interior and exterior glass, wetting, pulling the water off and polishing edges.

The egocentric window cleaning page covers the same two catalog rows and describes the capture as first-person video with ten streams, delivered in RLDS, LeRobot or WebDataset. The stereo floor mopping page and the stereo floor stripping and waxing page cover floor-care rows and do not list the window rows.

This page is the stereo and MCAP treatment of just the two window and glass rows. It does not repeat the catalog narrative of the egocentric page or any floor steps. The comparison table below sets the pages side by side.

Window cleaning is about the state of a pane: where it has been wetted, where the water has been pulled off, which edge is left to polish. A stereo pair recovers depth from the calibrated frame pair rather than from a standalone depth sensor, so the pane and the tool share a frame with the hands. The site gives no depth accuracy figure and no label for how clean a pane is, so whether wet or cleaned glass is resolvable in depth is Not stated.

The hard-environments page states that stereo keeps working outdoors where active depth sensors struggle, and that many programs capture both and sync them on one clock. The catalog rows list interior and exterior glass, but the site does not say where any batch is captured, so the setting is agreed in the spec.

The site describes two stereo setups and does not say which one a window 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 window 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.

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 glass work the site states hand pose, a stereo-derived depth stream, IMU, contact events, action segments and per-frame instance masks elsewhere. 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.

Hands are the first risk. A squeegee handle sits in the fist and an applicator or pole can hide the fingers, so the contact point is often hidden. The visibility flag on each joint is the answer the site gives, and contact events mark when a hand or tool meets an item. A spec can also ask for wrist views, so a grip the head camera loses is still recorded.

Depth is the second. The site says depth degrades on wet and reflective surfaces and masks low-confidence regions with per-frame confidence maps, and an episode is rejected when more than 35 percent of depth pixels in the manipulation region are invalid. A freshly wetted window is a wet surface, but the site does not say how depth behaves on glass, so this page promises no figure.

Calibration is the third, because 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 the manipulated object is out of frame for more than 15 percent of it or an undocumented bystander appears, and a reject log ships with every batch.

Commercial buildings are shared spaces, so bystanders matter, and glass can show people on the other side. 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. Blurring of signage or text seen through glass is not stated, so raise any such need before the spec is signed.

The site states no contributor safety protocol, no height, rope-access, ladder or lift setup, no guidance on cleaning products and no certification for any window row, so this page describes none. Capture at your own facility is listed as not offered.

Off-the-shelf covers what is already in stock: the stereo category lists the stereo datasets and hand-tracking video, and only dataset 06 is stated as MCAP. This page does not claim off-the-shelf coverage of the window rows. Custom collection covers the rest: custom collection at scale through a network of thousands of capture providers, for a consistent, recurring data flow to labs.

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. The site does not say which applies to these rows, so ask when you request the quote. Collections are billed per validated hour.

Describe the tool, the glass, the setting and the steps in order, what counts as one completed pass, 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.