Egocentric Produce Packing and Grading Video Data

Egocentric produce packing and grading video data is real-world, first-person video of produce handling in a packhouse or on a farm. It is recorded from the worker viewpoint, with depth, 3D hand pose, contact events and frame-accurate action segments aligned on one hardware clock. The tasks it covers are the ones the site states in its task catalog: sorting and cleaning strawberries, weighing, packing, sealing and labelling them into clamshell containers, packing graded produce into crates, and grading crops by size and quality.

The packing and kitting page covers packing work in general, and the sorting task page covers sorting in general. Neither is about fruit and vegetables. This page covers the produce sequence, from the first item picked off a sorting table to a labelled, closed container.

The site states produce packing in its task catalog, in a Produce Packing Facility environment under the Agriculture & Grounds sector. The environment lists three rows, all about fresh strawberries, and its examples are a farm packhouse with sorting tables, weighing stations and packing lines.

Three more rows state produce grading or packing in two other agriculture environments: Crop Sorting & Grading under Farm & Field Operations, and Pack produce crate and Sort + grade produce under Greenhouse, Nursery & Garden Centre. This page uses all six rows and lists their skill rating and status as the catalog states them.

The catalog row reads "Inspect, clean, and sort fresh strawberries, removing stems and damaged fruit, using sorting trays and collection bins." Its description is a strawberry farm processing area with sorting tables and washing stations. It is the stated inspection and culling task, and it comes before any weighing or packing.

The row is rated Easy and is Available on request. It names strawberries, stems and damaged fruit. The site does not state what counts as damaged, so the defect rule is set in the written spec.

The catalog row reads "Sort, weigh, pack, and label fresh strawberries into clamshell containers using a bench scale and a label gun." Its description is a produce packing facility with sorting tables, weighing stations and a dedicated packing line. One row therefore covers the whole run from sorting to a labelled container.

The row is rated Easy and is Available on request. Clamshell sizes, pack counts and target weights are Not stated, so they are set in the spec.

The catalog row reads "Weigh, pack, seal, and label fresh strawberries into clamshell containers using a bench scale and a label gun." It starts at weighing, so it leaves out the sorting step, and it adds sealing the container before the label goes on.

The row is rated Easy and is Available on request. The site does not describe a scale reading or a label as a recorded field, so neither is claimed here.

The Farm & Field Operations environment lists a Crop Sorting & Grading row that reads "Sort harvested crops by size and quality, discard damaged items and grade into crates." It is the stated grading task, and it names crops in general, not one fruit.

The row is rated Easy and is listed as Ready to deploy. Grade classes and grading standards are Not stated, so the spec sets the classes and what each one means.

The Greenhouse, Nursery & Garden Centre environment lists two produce rows. Pack produce crate reads "Pack graded produce into crates," and Sort + grade produce reads "Grade produce by size and quality." Both are rated Easy and Available on request.

The site does not state why their capture mode differs from the packing rows, so choose the mode in the spec.

A Field Crate Loading & Hauling row fills, stacks and hauls field crates, a Harvest and sort produce row cuts produce with shears, and a retail Produce handling row unpacks crates, trims and culls items and stacks displays. They name produce, but they are hauling, harvesting and shelf work, so they are listed as neighbours and not counted.

Generic packing, kitting, sorting, parcel and grocery put-away lines on the site do not name fruit, vegetables or produce. They have their own pages and are not counted here.

The catalog places the packing rows in a farm packhouse and the grading rows on a farm and in a greenhouse or grow-bench set. The site names no farm, packhouse, city or region, and it does not state capture at a customer own farm, packhouse or facility.

Environment, access and schedule are agreed in the spec. The number of facilities and the seasons covered are Not stated.

A head-mounted phone records the first-person view of real work on site, and a multiview rig adds synchronized third-person angles when the spec asks for them. The greenhouse rows list a multiview rig with handheld grippers instead.

The site does not state a camera count, resolution or frame rate for produce work. Line speed and throughput are Not stated, so session length and volume are quoted against your spec.

Each episode follows the Firsthand episode schema: ten streams synced under 2 ms, with head and wrist RGB, depth, 3D hand pose with per-joint visibility, contact and grasp events, and frame-accurate action segments. For produce, hand pose tracks how the fingers pick and place each item, contact events mark when fingers meet fruit, a tray or a container, and action segments mark each step.

Object labels use the terms your spec sets. Force or tactile data, and defect or quality-outcome labels, are Not stated. They are defined in the spec if your policy needs them.

Delivery formats are RLDS, LeRobot and WebDataset, plus HDF5, zarr and Rerun .rrd. The format table below sets each against what its official documentation states.

Every batch is validated against signed accept criteria before it ships, with a reject log, and delivery is rolling, per validated batch.

Every episode links to a signed contributor release through a documented consent chain, and the consent record is part of the delivery. Faces, plates and screens are detected and irreversibly blurred at ingest, and a reviewer confirms the result before anything ships. The anonymization record is delivered with the data.

The site states no further privacy rule specific to packhouses or farms, so this page makes none. The consent chain, the blur pass and the anonymization record are the stated privacy steps.

Where coverage is already deep, the first validated batch lands in a median of 48 hours from signed spec. Thin environments typically take three to four weeks, because vetting operators there is the bottleneck, and we quote that up front.

On-demand top-ups run against the same spec and accept criteria after each training run. Most produce rows are Available on request, so the site states no capture timeline for them beyond a quote against your spec.

You describe the produce, steps and sequence your policy needs. We turn that into a written spec covering tasks, environments, capture tier, annotation depth and accept criteria, then quote reach, timeline and price before any capture begins.

A Spec Pilot, a small and fast first batch, can prove the spec and accept criteria before you scale. Billing is per validated hour, a reject log ships with every batch, rejected items are recaptured at our cost, and every collection ships under a perpetual, buyer-owned commercial license.