Unlocking Mixed Depalletizing Challenges: How Transfer Technology's 3D Vision Achieves Precise Picking of 7 Types of Cartons?

In intelligent warehousing and production line logistics, carton depalletizing remains a highly labor-intensive process. Traditional mechanical depalletizing solutions struggle with mixed pallets featuring cartons of varying sizes and bag-wrapped surfaces, while manual handling is constantly limited by low efficiency and inconsistent operating quality.

To tackle these industry pain points, Transfer Technology has deployed a reliable vision-guided depalletizing system powered by the Epic Eye Laser L 3D smart camera in a large-scale manufacturing project, delivering a stable and efficient automated solution for complex mixed depalletizing scenarios.

Project Challenges

The project requires the vision system to recognize and process 7 different carton specifications, among which 3 types are wrapped in outer packaging bags. The comprehensive application challenges are summarized as follows:

1.Wide dimensional variation

Carton sizes range drastically from 150×150 mm (minimum) to 600×400 mm (maximum), requiring strong adaptive recognition capabilities.

2.Complex surface interference

Packaging bags feature reflective surfaces, soft textures and irregular wrinkles, which easily cause recognition instability for conventional machine vision systems.

3.Irregular mixed stacking patterns

Incoming pallets adopt mixed stacking modes with unconstrained carton positions and orientations, eliminating fixed picking rules.

4.Strict cycle time requirements

Continuous high-volume production demands ultra-fast vision processing and highly coordinated robot motion to meet high-takt operational standards.

Core Technology

Deployed as the core visual perception unit, Transfer Technology’s Epic Eye Laser L 3D camera supports the full set of mixed depalletizing logic and builds a robust vision-guided robotic depalletizing solution.

Adopting professional laser 3D imaging technology, the camera delivers outstanding industrial adaptability:

Hgh-precision point cloud reconstruction

It generates complete, high-resolution 3D point cloud data with clear contours, even under complex interference such as bag reflection and inconsistent carton colors.

Strong anti-interference capability

Laser active imaging is barely affected by ambient light fluctuations, ensuring stable recognition under harsh on-site lighting conditions.

Ultra-fast processing speed

The entire scanning and computing process is completed within 4 seconds, fully matching high-speed production cycle requirements.

Solution Overview

The system realizes an intelligent closed-loop workflow of perception, decision-making and execution through coordinated vision and robotic control.

1.Intelligent Recognition & Positioning

Mounted on a mobile gantry, the 3D camera alternately covers two feeding stations. It processes real-time point cloud data to accurately identify carton types, XY coordinates, height and rotation angles for both bare and bag-wrapped cartons, and transmits precise pose data to the robot.

2.Adaptive Gripping Strategy

Based on real-time visual feedback, the robot automatically switches suction zones on the vacuum gripper to fit different carton dimensions. For bag-wrapped cartons, dedicated algorithm compensation effectively prevents slipping and ensures stable and reliable gripping.

3.Dual-station Cyclic Operation

Once depalletizing is finished at Station A, the camera automatically shifts to Station B while triggering material replenishment at Station A. This dual-station alternating mode enables nearly non-stop depalletizing, maximizes robot utilization, and minimizes manual intervention.

Customer Value

Verified by on-site commissioning and continuous operation, the system achieves outstanding industrial-grade performance:

99.9% High Recognition Accuracy

Maintains a comprehensive recognition accuracy of over 99.9% for all 7 carton types under standard incoming conditions.

4-second Ultra-fast Cycle Time

Stably completes full-scan positioning and pose calculation within 4 seconds per cycle to support high-efficiency production.

Long-term Operational Stability

The system successfully passed an 8-hour non-stop continuous operation test with zero failure, fully proving its reliability in mass production scenarios.

As vision hardware and algorithms continue to iterate, 3D vision technology will empower robots to handle increasingly complex industrial scenarios. Transfer Technology will keep optimizing core 3D vision perception technologies and deliver more stable, efficient and cost-effective intelligent solutions, empowering automated upgrading for the manufacturing and logistics industries.