Case Study | Tote Depalletizing and Palletizing: 3D Vision Enables Millimeter-Level Stable Operations

Tote depalletizing and palletizing are core high‑frequency procedures in warehouse inbound and outbound automation. Traditional manual operations are plagued by persistent efficiency bottlenecks and high labor intensity. Meanwhile, rigid mechanical positioning solutions feature poor scene adaptability. Faced with multiple tote specifications and variable stack heights, such solutions frequently suffer from positioning deviations and grasping failures, severely restricting production line efficiency improvement.

Recently, Transfer Tech deployed a 3D vision intelligent loading and unloading system for a domestic power industry warehouse. Powered by the self‑developed Epic Eye Pixel Pro 3D industrial camera, the system enables fully automated tote depalletizing and palletizing with consistent millimeter‑level positioning accuracy.

01 Project On‑Site Conditions

This project adopts a single‑station layout, where one 3D vision system pairs with a single industrial robot to handle both depalletizing and palletizing workflows. For depalletizing, fully stacked totes are delivered in order via material carts; for palletizing, individual totes are continuously supplied through conveyor lines. The system is designed to handle two types of standard plastic totes with different heights.

The on‑site scenario presents three core challenges: First, variable incoming stack heights eliminate fixed robot motion paths, requiring real‑time visual positioning for every operation.

Second, two tote sizes share one working station, demanding strong and rapid compatibility of the vision system.

Third, the production line requires high continuous operational stability, with extremely low tolerance for recognition errors and grasping failures.

02 Solution

To adapt to the on‑site working conditions, an eye‑in‑hand configuration is adopted. The Epic Eye Pixel Pro 3D industrial camera is mounted on the robot end‑effector and moves synchronously with the robotic arm. This configuration removes the need for fixed top‑mounted brackets, delivers superior spatial adaptability, and perfectly accommodates scenarios with dynamically changing stack heights.

During operation, the robot carries the camera to scan the top of the tote stack. The 3D vision system rapidly calculates the position and posture of the uppermost tote, transmits coordinate data to the robot controller, and guides the robotic arm to grasp and place totes precisely onto the conveyor line.

The robot automatically adjusts the scanning height according to unknown stack variations, ensuring the camera always works within its optimal depth range. Furthermore, the system supports synchronous material model verification during grasping, validating internal material specifications to eliminate material mixing errors at the source.

03 Core Performance

Verified by on‑site joint commissioning and testing, all key performance indicators of the integrated vision system fully meet on‑site requirements and support stable long‑term production line operation.

Controllable Positioning Accuracy: The tote positioning accuracy reaches ±2 mm, ensuring high‑precision end grasping and effectively avoiding tote scratching, offset placement and other operational anomalies.

Stable and Reliable Recognition: The system achieves a comprehensive recognition rate of ≥99.9%. Rare recognition failures occur under standard working conditions, minimizing manual intervention and enabling uninterrupted long‑term line operation.

High Operational Efficiency: The full cycle of single scanning and algorithm processing takes ≤3 seconds, fully matching the cycle requirements of mainstream warehouse inbound and outbound stations.

This outstanding performance is supported by Transfer Tech’s self‑developed Epic Eye Pixel Pro 3D industrial camera. Built on structured light imaging technology, the camera features a 2.3‑megapixel resolution and a maximum measurement accuracy of 0.1 mm, enabling clear reconstruction of tote edges and surface features. It maintains stable feature matching even when totes have minor deformation within 2 mm.

Transfer Tech has developed mature standardized products and turnkey solutions for typical industrial scenarios including palletizing, depalletizing, and machine tending. Supported by fully self‑developed hardware and algorithm technologies, the company can rapidly deliver customized solutions for diverse industries and provide reliable visual perception support for automated production lines.