Named steps
Clearing debris, sorting waste, sweeping, loading a wheelbarrow and hauling to the skip are different steps, so a spec names each one and the order they run in.
Task family
Requirements
Clearing debris, sorting waste, sweeping, loading a wheelbarrow and hauling to the skip are different steps, so a spec names each one and the order they run in.
The catalog rows say debris, waste, brooms, wheelbarrows and a site skip, and no more, so the spec says which materials and tools your policy needs.
A head-mounted view or a multiview rig keeps the hands, the broom or wheelbarrow handles and the load in frame through the whole sequence.
Frame-accurate action segments mark where sweeping ends, where loading starts and where a haul is finished against the accept criteria in the spec.
A perpetual, buyer-owned commercial license backed by a documented consent chain.
Why it matters
Egocentric construction site cleanup data is real-world video of the physical work of leaving a site clear after works. A worker clears debris, sorts waste, sweeps it up, loads it into a wheelbarrow and hauls it to the site skip. It is recorded against a written spec, with the streams the site states for its custom collections.
The trash and recycling page and the sorting page list these rows inside larger tables, and the floor cleaning page lists one of them. The construction and skilled trades page covers jobsite trades more broadly. This page centres the two cleanup rows. It describes task video only, and it makes no safety, OSHA, hazardous-material or certification claim.
Catalog rows
Two, both in the Active Construction Site environment under the Construction & Trades sector. "Construction Site Cleanup & Debris Removal" is described as clearing debris, sorting waste and preparing sites after works. It is rated Easy and listed as Active collection.
"Sweep and haul construction debris using brooms and wheelbarrows" is described as sweeping up debris, loading it into wheelbarrows and hauling it to the site skip. It is rated Easy and listed as Active collection. Both rows list a head-mounted phone or a multiview rig as the capture mode, and the steps inside each task are set by your spec.
The difference
The Active Construction Site environment has 15 rows, and two of them are cleanup rows. The others, such as carpentry, plumbing, tiling, drywall, insulation and painting, are separate listings. The trash and recycling page covers waste handling more broadly, and the floor cleaning and mopping page covers floor work.
For jobsite trades, see the construction and skilled trades page. For waste handling, see the trash and recycling page. For tool handling across settings, see the tool use page.
Streams
Firsthand custom collections record ten hardware-synchronized streams aligned under 2 ms, with head and wrist RGB, depth, 3D hand pose with a per-joint visibility flag, contact and grasp events and frame-accurate action segments. The one language label the site states is a task instruction for each episode.
Standard capture is cameras, depth and IMU. A separate glove tier records a tactile signal, and the site does not state whether it applies to these rows. The site lists no label for the outcome of a step, such as whether a site is clear. Debris types, wheelbarrow and broom models and whether force or tactile data are needed are settled in the spec.
Privacy
Every episode links to a signed contributor release through a documented consent chain, and the consent record is part of the delivery. Faces and screens are detected and irreversibly blurred at ingest, and a reviewer confirms the result before anything ships. A shared jobsite can put other trades and crew in frame, and a clip where someone cannot be de-identified is discarded.
That is the whole of the privacy wording the site states. It does not state a rule for specific construction sites or contractors, and the site states that Firsthand does not capture at a customer facility. Which site a program uses is agreed in the spec.
Fast · on demand
Where coverage is already deep, the first validated batch lands in a median of 48 hours from signed spec. The construction and skilled trades page describes construction as a thinner-coverage setting, where a first batch typically takes three to four weeks, because vetting operators with site access is the bottleneck, and we quote that up front. The site does not state which of the two applies to these two rows.
You describe the debris, tools, hauling route and sequence your policy needs, and we turn that into a written spec covering tasks, environments, capture tier, annotation depth and accept criteria. A Spec Pilot, a small and fast first batch, can prove the spec before you scale. Billing is per validated hour, a reject log ships with every batch, and rejected items are recaptured at our cost.
At a glance
Every collection runs through the same seven-stage end-to-end custom collection pipeline. End-to-end custom data collection is a managed service that takes an AI data need from problem to owned dataset in a single accountable pipeline.
Catalog rows
| Catalog task | Environment and sector | Catalog description | Skill | Status |
|---|---|---|---|---|
| Construction Site Cleanup & Debris Removal | Active Construction Site, Construction & Trades | Clearing debris, sorting waste and preparing sites after works | Easy | Active collection |
| Sweep and haul construction debris using brooms and wheelbarrows | Active Construction Site, Construction & Trades | Sweep up debris, load it into wheelbarrows and haul it to the site skip | Easy | Active collection |
Streams and tasks
| Stream or label | Status on this site | What it carries in a cleanup episode |
|---|---|---|
| Head and wrist RGB, depth | Stated; ten streams synced under 2 ms | Keeps the debris, the tool and the hands on one timeline |
| 3D hand pose, per-joint visibility | Stated for custom collections | Carries how fingers hold a broom, a handle or a load |
| Contact and grasp events | Stated | Marks when hands meet a broom, a wheelbarrow or a piece of debris |
| Action segments | Stated; frame-accurate | Separates clearing, sorting, sweeping, loading and hauling |
| Task instruction | Stated; one per episode | States the cleanup task in plain language |
| Object labels | Set by your spec | Names the debris, tools and steps in the terms your policy uses |
What is not stated
| Item | Status on this site | Where it is settled |
|---|---|---|
| Safety, OSHA or hazardous-material claims | Not stated | Not claimed on this page |
| Certification claims | Not stated | Not claimed on this page |
| Debris types, waste categories and how waste is sorted | Not stated; the rows say debris and waste only | The written spec |
| Broom, wheelbarrow and skip models or sizes | Not stated; the rows say brooms, wheelbarrows and a site skip only | The written spec |
| Site type, trade or stage of works | Not stated beyond the Active Construction Site environment | The written spec |
| Outcome labels such as site clear or skip full | Not stated | The written spec |
| Facility names, regions, number of sites or workers | Not stated beyond Western program markets | The written spec |
| Weights, volumes, cycle times, throughput, repetition counts | Not stated | Quoted against your spec |
| Force or tactile data | Not stated for these rows | The written spec, if needed |
| Robot or policy results | Not stated | Not claimed on this page |
Formats
| Format | What the official documentation states | Used for here |
|---|---|---|
| RLDS | A dataset is retrieved as episodes, and each episode holds a dataset of steps (RLDS repository README) | Episode-based training loaders |
| LeRobot | LeRobotDataset v3.0 is a standardized format for robot learning data, including multi-camera video, with many episodes stored per Parquet or MP4 file (LeRobot documentation) | LeRobot dataset format and tooling |
| WebDataset | Tar files in which files that belong together share a basename, split into numbered shards (WebDataset README) | Sharded streaming for large training runs |
| HDF5 and zarr | Stated by Firsthand only | Array storage for custom pipelines |
| Rerun .rrd | Stated by Firsthand only | Multi-stream synchronized playback for review |
FAQ
Sources
Tell us the debris, tools, hauling route and sequence your policy needs to handle. We will quote reach, timeline, and price against a written spec, with a first validated batch in a median of 48 hours where coverage is deep.