Stereo Hedge Trimming and Pruning Data

It is first-person video of a person trimming and pruning a hedge or plant, recorded with two calibrated cameras and with hand pose and sensor streams alongside. The tasks are the two rows the site’s catalog lists in Farm & Field Operations. The first reads "Prune and trim hedges using hedge trimmers and loppers" and is described as cutting hedge tops and sides with trimmers, removing thick stems with loppers and clearing cuttings. The second is named "Pruning" and is described as cutting back dead and overgrown branches with shears and loppers to shape the plant. Stereo describes the capture, and MCAP describes the container. They are separate choices.

The data is human capture by paid, consented contributors, not robot or teleoperation data. No named model was trained on Firsthand data for this page, and Firsthand is not affiliated with Foxglove or the MCAP maintainers.

Two, both in Farm & Field Operations under Agriculture & Grounds, in the catalog data file lib/commercial-catalog-data.ts. The environment example line reads "Farm fieldwork, greenhouse, orchard, vineyard -- harvesting, planting, pruning, irrigation, equipment operation". The task "Prune and trim hedges using hedge trimmers and loppers" has the description "Cut hedge tops and sides with trimmers, remove thick stems with loppers and clear cuttings.", is rated Medium, lists "Phone (head-mounted) / multiview rig" as its capture setup and is listed as Active collection.

The task "Pruning" has the description "Cut back dead and overgrown branches with shears and loppers to shape the plant.", is rated Medium, lists the same capture setup and is also listed as Active collection. Both rows were read on the live catalog and the Farm & Field Operations environment page. The first table below reproduces them as the catalog states them, and the site states no hedge species, hedge height, hedge length or tool model for either row.

The site defines Active collection as recording that is running now, so a spec can join an existing capture stream. That differs from Ready to deploy, which means contributors and site access are set up and capture starts once a spec is signed, and from Available on request, which means the work is scoped and recruited to spec when a buyer asks for it. Both hedge rows carry the first label.

The site states no delivered hours, no clip count and no delivered stereo footage for either row, so this page claims none. What an order delivers, and when, is quoted against your spec. Active collection describes the catalog capture setup, which is a head-mounted phone or a multiview rig. It does not describe stereo.

Not stated. The catalog lists a head-mounted phone or a multiview rig as the capture setup on both rows, and says nothing about a stereo pair. Stereo applies only through the site’s general stereo offer: a stereo pair is a capture choice agreed in the skill spec and confirmed on a Spec Pilot. This page therefore describes stereo as an option for an order, never as the setup the catalog lists.

That distinction matters when you plan a program. If you want a calibrated pair, you say so in the spec, and the site’s stated stereo rigs are the reference for what that can mean. If you are content with head-mounted phone capture or a multiview rig, the Active collection rows are the closest match to what is recorded now.

Four pages already touch this work, and none has hedge trimming as its main topic. The egocentric landscaping and groundskeeping page covers three grounds rows, with the hedge row as one of them beside commercial and farm grounds, and treats capture as egocentric. The stereo landscaping and groundskeeping page is the stereo and MCAP treatment of the one commercial grounds row, and states that the hedge row sits on the egocentric page instead.

The egocentric pruning and trellis page covers five rows, four on a greenhouse bench plus the farm row "Pruning", and does not cover the hedge row. The stereo greenhouse potting and pruning page covers seven greenhouse bench rows and treats "Pruning" as context only. The egocentric farm and field task page lists the hedge row for orientation among 21 farm rows. This page is the stereo and MCAP treatment of the two rows that name hedge trimmers, loppers and pruning together. The second table below sets them side by side.

Hedge work is about where the blade edge sits relative to the stems and the hedge face: how far the trimmer is from the surface, where a lopper jaw closes on a thick stem, where the cuttings fall. A stereo pair recovers depth from the calibrated frame pair rather than from a standalone depth sensor, so the tool, the plant and the hands share a frame.

The site gives no depth accuracy figure and no label for a straight cut, an even hedge face, a clean stem removal or a cleared path, so whether those outcomes are resolvable in depth is Not stated. Capture at your own facility is listed as not offered, and the site names no grounds crew, farm or site for either row.

The catalog states one capture setup for both rows: a head-mounted phone or a multiview rig. A head-mounted phone records the first-person view, 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 hedge 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 hedge or pruning 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. The site does not state outdoor lighting, weather, season or terrain for these rows, so those are Not stated and are agreed in the skill spec.

The site states that 3D hand pose is delivered as two hands of 21 joints each in metric units in the head-camera frame, with a visibility flag per joint, aligned to depth and RGB. For hedge work, that records how the fingers hold a trimmer or a lopper, and the flag records which joints were hidden behind a tool or foliage. Contact events mark when fingers or a tool meet a plant, and action segments mark where cutting tops and sides ends and removing thick stems begins.

The site states that egocentric video puts hands entering the frame from below, occlusion during grasp and motion blur during reach into the training distribution. A trimmer or a pair of loppers held out from the body is exactly that case. Force or tactile data is Not stated for these rows. The site offers a tactile glove tier as a pilot-ready option, but does not state that it has been used on a hedge or pruning row, so treat force data as an add-on to agree 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 hedge and pruning work the site states hand pose, depth recovered from the stereo pair, 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. The only language data tied to Firsthand capture is one task instruction per episode.

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.

Hand occlusion is the first risk. The site states that fingers disappear behind objects, and a hand closing loppers on a thick stem or gripping a trimmer behind foliage is partly hidden from the head camera. The visibility flag on each joint is the answer the site gives for hidden fingers, and a spec can ask for wrist views so a grip the head camera loses is still recorded.

Cluttered, reflective and moving scenes are the second. The site names reflective surfaces and clutter as cases where a single camera struggles in hard environments, and notes that wet and reflective surfaces degrade active-stereo depth in its kitchen dataset. Whether thin stems, dense foliage, wet leaves, cut branches on the ground or direct sun behave the same on a stereo pair is Not stated. The site ships per-frame depth confidence maps for its kitchen captures, and whether a hedge order ships them is Not stated, so ask in the spec.

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. Force and tactile readings are not among the listed streams, and the site states no stem, leaf or cut label, so what counts as one cut or one cleared pile is written into the spec.

A hedge is trimmed in passes that look alike on camera: one length of the top, then the next, one side, then the other, one thick stem removed and then another. Action segments carry start, end, verb and noun, so the spec decides where one segment ends and the next begins, and which object is the noun.

The site states no episode length, no segment count and no per-cut label for these rows, and the catalog does not split either row into finer steps beyond the three the first description names: cutting tops and sides, removing thick stems and clearing cuttings. Whether an episode is one pass, one hedge section, one plant or a full hedge is set per program. Put that boundary and the verbs you expect into the skill spec, and check them on the Spec Pilot before you scale.

Episodes are machine-checked at ingest against the stated reject criteria, and a human reviewer then checks the result. The site states these reject criteria for every episode: the manipulated object out of frame for more than 15 percent of the episode, more than 35 percent invalid depth pixels in the manipulation region, a sync breach beyond the ceiling for more than three consecutive frames, calibration drift beyond tolerance, an undocumented bystander, and a mismatch against the named skill spec.

The site states no blur or sharpness rule and does not state how often a plant or tool leaves the view outdoors. A rejected item goes to the reject log, which ships with every batch, and is recaptured at our cost. Each batch is validated against signed accept criteria before it ships.

For every delivered episode the site states that faces, plates and screens are detected and irreversibly blurred at ingest, that a reviewer confirms the result before anything ships, that any identifiable un-consented person goes to human review, and that footage that cannot be de-identified is discarded. Every episode links to a signed contributor release through a documented consent chain, and the consent record and the anonymization record are delivered with the data.

Hedges often border roads, paths and neighbouring property, and who may be in frame is agreed in the spec. The site states no plant species, hedge size, site name, crew, task time, throughput or error rate for these rows, and each is Not stated. This page makes no safety, certification or tool-handling claim beyond the site’s own wording.

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 an off-the-shelf dataset of hedge trimming or pruning, so this is custom collection: 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 speed applies to a hedge order, so ask when you request the quote. Collections are billed per validated hour, and on-demand top-ups run against the same spec and accept criteria after each training run.

Describe the hedge or plant work and its order, the tools involved, what counts as a finished pass, the setting 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.