Egocentric Drilling and Wall Mounting Video Data

Egocentric drilling and mounting video data is first-person video of a person drilling a hole, then mounting, hanging or installing something into or onto a workpiece, recorded from their own point of view with synchronized sensor streams alongside. The camera sits on the head, close to what a robot’s onboard camera would see at a bench or on a job, so the frames show how the hands hold the tool, line it up with the work and move on to the fixing. Teams searching for drilling robot training data, a power drill video dataset or egocentric mounting and installation data are usually looking for this.

On this site the data is human capture: a paid, consented contributor wears the rig and performs a task from a written skill spec. It is not robot or teleoperation data, and it is not a fixed corpus you download. It is custom installation task data collection, scoped to your tasks. This page covers a different subject from screwdriving and fastening data, which centres on driving and tightening fasteners, from tool use data, which centres on handling tools, from furniture assembly data, which covers putting units together, and from construction and trades data, which covers whole jobs.

The task catalog names drilling, mounting or installation steps in five environments, ten rows in all. Five rows state drilling. The Carpentry & Furniture Workshop lists three, all Medium: assembling and joining furniture frames, which glues and clamps frame joints, drills pilot holes and screws the frame together; drilling and installing cabinet hardware with a hinge jig, which positions hinges, drills pilot holes and screws hinges and handles onto cabinets; and "Hardware Installation", which marks, drills and fits hinges, handles and slides onto cabinets and furniture. Two of these are Active collection and one is Ready to deploy.

Heavy Industry & Fabrication lists "Metal fabrication", rated Hard and Available on request, which cuts, bends, drills and welds metal parts into assemblies using shop machinery. The Trades Workshop Bench lists "Drill pilot + countersink", which drills a pilot hole and countersinks, rated Medium and Available on request, and "Hang a fixture", which levels, marks and screw-mounts a fixture. Both bench rows are recorded with a multiview rig and handheld grippers rather than a head-mounted phone.

Four rows mount or install without a stated drilling step. On the Active Construction Site, "Electrical installation" runs cable, mounts boxes and wires outlets, switches and fixtures, and "Plumbing" cuts, joins and installs pipes and fixtures and tests for leaks; both are rated Hard and Active collection. The Data Center & IT Hardware Bench lists "Clip cage nuts + panel", which clips cage nuts and mounts a blanking panel, and "Screw-mount rack rails", which fastens rail brackets with a driver. Both are Available on request and use handheld grippers.

The first table below lists every row. Rows that only drive or tighten fasteners, tire and wheel mounting, and rows where "mount" means placing a work blank on a machine are left out, and the screwdriving and tire pages cover those. No row states a wall, a drill type, a bit size, a torque value or a load rating, so whether a task mounts onto a wall, a panel or a frame is Not stated and is written into the skill spec.

Each attempt is delivered as one episode. On this site the standard episode carries action segments, 3D hand pose for both hands with a per-joint visibility flag, contact events that mark when a hand meets an object, an objects list, and per-frame instance masks stored at the depth resolution of 848×480. The capture setup table below lists the streams and sync figures the site states for the instrumented tier: ten streams hardware-triggered onto one PTP clock domain and aligned under 2 ms, 6-DoF IMU data at 200 Hz, and calibration twice per session.

The catalog lists a head-mounted phone with a multiview rig for six of these rows, and a multiview rig with handheld grippers for the other four, while the ten-stream stack is the instrumented tier. Which applies to a drilling program is not stated and is confirmed in the skill spec. Force and tactile readings are not in standard capture, and whether a glove tier applies to a drilling or mounting program is Not stated.

Action segments divide an episode into start- and end-timestamped spans, each carrying start_ns, end_ns, verb and noun. For drilling and mounting work, a segment might be one mark, one hole drilled, one hinge positioned, one screw driven or one fixture hung, as written in the skill spec. Where one completed step starts and ends is a spec decision, not a default, and so is whether drilling and fixing are one segment or two.

Validation runs against signed accept criteria written into the spec before capture, and a reject log ships with each batch. Episodes are machine-checked against stated reject criteria, including a sync breach, calibration drift, the manipulated object out of frame for more than 15 percent of the episode, an undocumented bystander, and a mismatch against the named skill spec, and a human reviewer then checks each episode. Only footage that clears the criteria is billed as a validated hour, and rejected items are recaptured at our cost.

Delivery is in the format your stack already reads: RLDS, LeRobot, WebDataset, HDF5, zarr or Rerun .rrd, or raw streams with metadata, with readable converter source alongside. The license is perpetual, irrevocable and worldwide, buyer-owned and non-exclusive by default, with exclusivity available as a separate term. The executed agreement controls.

Drilling and mounting combine a tool, a surface and a fixing in a few seconds of close-range motion. The drill and the hand often cover the exact spot being worked, so the fingers and the tool hide the hole from the camera, which is why hand pose carries a per-joint visibility flag, contact events mark when the hand meets the object and masks are recorded per frame. Two hands usually work at once, one holding the piece, the level or the jig while the other runs the tool, and a model has to learn that coordination as well as the order of mark, drill and fix. Deciding when a hole is deep enough or a screw is seated is a judgment, and the site states no depth, force or success label. The second table below maps each challenge to the stream or label this site states, and says so where nothing is stated.

A robot foundation model paper on arXiv organizes its training data as a pyramid, with large quantities of web data and human videos at the base, synthetic data in the middle and real robot data at the top. The same paper says human videos have no explicit action labels, so it learns latent actions to label them. That is robot-model research, not Firsthand data, and Firsthand has no relationship with its authors.

The catalog describes the head-mounted phone as recording the first-person view of real work on site. The site also states that Firsthand does not capture at a customer’s own facility. For the four handheld-gripper rows, the site describes the mode as a staged working set recorded while the operator uses the grippers. For the other six rows, whether a program uses a working site or a staged set is not stated, and neither is the venue type. Both are agreed in the skill spec before capture begins.

Every contributor signs a written release, and each delivered episode links to it through consent artefact IDs in the batch data card. The consent chain has six checkpoints, and a break at any one discards the footage. At ingest, automated detection finds faces, plates and screens and blurs them irreversibly. Speech is scrubbed of names, addresses and card numbers, while ambient audio is kept. A human reviewer confirms that no un-consented person is identifiable in any frame, footage that cannot be de-identified is discarded, and the anonymization record ships with the data.

The site states no safety, tool-handling or certification practice for these environments, so this page makes no such claim.

From signed spec to first validated batch, the published timeline is a median of 48 hours where coverage is deep and three to four weeks where it is thin and contributors need to be recruited first. Four rows are Active collection, where recording is running now and a spec can join an existing capture stream, and one is Ready to deploy, with contributors and site access set up. Five are Available on request, scoped and recruited to spec. The range that applies is confirmed when the spec is quoted, which is what makes fast, on-demand custom egocentric data collection possible for a drilling or mounting task that no stock dataset covers.

Describe the drilling and mounting tasks and their order, what counts as a completed hole, fixing or installed part, who may be in frame, and the variation cases your pipeline is missing. That becomes a written skill spec, quoted for reach, timeline and price before any capture begins. Our skill spec guide walks through how to write one.

Most programs begin with a Spec Pilot, a small, fast first batch that proves the spec and the accept criteria on real footage before volume scales. Collections are billed per validated hour. Counts of sites, contributors, episodes and hours are set in the skill spec, since the site does not state them for these environments.