3D Industrial Camera Enables Automated Carton Depalletizing: Flexible and Efficient, Solving Complex Scene Challenges

In the process of automation upgrade in manufacturing and logistics warehousing, carton depalletizing is a core link in material flow. With the popularization of flexible production modes, the same depalletizing station needs to accommodate multiple carton specifications, multiple stack patterns, as well as the identification and sorting of abnormal cartons. Traditional methods relying solely on conveyor line positioning or mechanical limits can no longer adapt to complex and changing field requirements. 3D vision-guided technology has thus become the mainstream solution in the industry. However, in actual implementation, visual positioning accuracy, operation cycle time, and adaptability to field environments remain core challenges limiting stable project operation.

I. Project Challenges

The incoming materials at this station are standardized cartons, each weighing about 20 kg. The cartons are tightly stacked in 2 rows and 4 columns on a pallet, forming two layers. The production line has two types of difficulties: 1. Complex and unfixed stack patterns Full pallet incoming: The number of cartons in the upper layer varies from 1 to 5, with a total of 9 to 13 cartons. Partial pallet incoming: Pallets re-entered after partial picking, with only a single layer or a few cartons on the upper layer, numbering 1 to 8 cartons. 2. "Zero-carton" diversion Some cartons have a white label ("zero-carton" mark) on the top surface due to incomplete filling. Zero-cartons must be placed by the robot onto a tail-off platform rather than being palletized onto the discharge pallet. The zero-carton is fixed as the last one in the layer, but the label is attached with angle and position deviations, making it difficult for traditional template matching to identify stably.

Additionally, there are no gaps between cartons, and the surfaces have strapping bands and printed patterns, which place high demands on the point cloud completeness of the 3D camera and the recognition algorithm.

II. Solution

To address the above challenges, the project adopts a vision-guided depalletizing solution using a 3D industrial camera and an industrial robot. An Epic Eye Log L 3D industrial camera is deployed on-site, fixedly installed directly above the depalletizing station.

The system flow is as follows: A forklift transports the carton stack to the depalletizing station and triggers an arrival signal. The robot sends a zero-carton detection signal, and the 3D camera takes an image and scans to confirm the presence of zero-cartons. The camera takes an image, and the algorithm identifies the arrangement of cartons and the position of zero-carton labels, outputting the grasping point pose and the number of cartons to grasp (1 or 2 at a time). The robot grasps according to the vision results, palletizes the cartons onto the discharge pallet, and places zero-cartons onto the tail-off platform. Repeat until the stack is empty or a specified number is reached; if the discharge demand is less than the total incoming cartons, zero-cartons are placed back onto the original pallet for return storage.

III. Application Advantages

Dual excellence in accuracy and cycle time Measured recognition accuracy is ±1mm, grasping accuracy is ±1mm, and the time for imaging plus algorithm is only 3 seconds, ensuring sufficient cycle margin.

Stable and reliable zero-carton recognition Through deep learning training, the system's recognition success rate for white labels under different lighting and tilt angles is ≥99.9%. Zero-cartons are accurately placed onto the tail-off platform after being grasped, avoiding mixing with normal palletizing.

Adaptive to partial pallets and complex arrangements The camera calculates the number of cartons per column (1 or 2) after each image and automatically plans the grasping point. In partial pallet scenarios, no manual re-teaching is needed; the vision system directly outputs the poses of remaining cartons.

IV. Implementation Value

Compatible with zero-cartons/partial pallets: Eliminates manual sorting and reduces labor intensity for field operators. Efficiency upgrade: Replaces manual depalletizing, significantly improving efficiency with stable and controllable cycle time, breaking through the efficiency bottleneck of traditional solutions and adapting to mass production needs. Flexible production changeover: When adding new carton types, no hardware replacement is needed through deep learning training. With the trend of industrial automation towards flexibility and intelligence, 3D industrial cameras will continue to deepen their application in core scenarios such as depalletizing, sorting, and loading/unloading. With technical advantages of high precision, high stability, and strong flexibility, they will become the core driving force for manufacturing automation upgrades.