News activity

Unlocking Full Automation in Automotive Manufacturing: AI + 3D Vision Transforms Shaft Handling

Time:2025-05-01
In the automotive parts manufacturing industry, automating the loading and unloading of precision components like metal shafts has long been challenged by disorganized placement and multi-spec adaptation. Recently, a leading company implemented Transfer Technology's AI + 3D Vision guidance system, leveraging high-precision 3D recognition and dynamic trajectory planning to enable fully automated picking of multiple shaft specifications and manage tray layers, driving the unmanned upgrade of their production line.


01 Customer Pain Points


The production line needs to handle three different shaft specifications, and the incoming material comes in bins that are randomly stacked. Traditional methods, relying on manual teaching and 2D vision systems, present three major bottlenecks:


  • Unable to identify overlapping parts:

    2D vision struggles to detect the spatial positions of stacked shafts, requiring manual sorting assistance.

  • Inefficient changeover:

    Switching between product specifications requires recalibration, interrupting the production flow. 

  • Layer transitions depend on manual labor:

    After picking a single layer of shafts, workers must manually remove the tray before continuing the process.


02 Solution


Transfer Technology, based on proprietary industrial 3D camera and AI algorithm, creates a seamless, full-process picking solution:


  • High-Precision 3D Scanning


The industrial 3D camera scans the entire bin with ±1mm precision, completing the imaging in just 1.67 seconds. The AI algorithm automatically separates different shaft specifications, adapting to challenging conditions like metallic surface reflections and edge blurring.


  • Multi-Target Collaborative Computation


In a single shot, the system can recognize 3 to 5 shafts, completing the algorithm mapping in 3.88 seconds. With 1.1 seconds of dynamic trajectory planning, the system ensures parallel processing between the robotic arm's movement and system calculations.


  • Smart Layer Management


By detecting changes in point cloud height, the system determines empty layers and automatically triggers tray retrieval commands. It supports dynamic compensation of the bin’s position, eliminating the need for manual repositioning.


03 Customer Benefits


After implementing the system, the customer’s production process achieved a triple upgrade:


  • Simplified Operation


Operators can simply select the shaft model via the front-end interface. The system automatically adjusts parameters and completes visual calibration, eliminating production line downtime caused by manual changeover adjustments.


  • Enhanced Precision and Reduced Manual Intervention


With ±1mm recognition accuracy and ±2mm picking precision, the system accurately locates overlapping shafts in the bin, even under interference from reflective metal surfaces, ensuring a high success rate for stable grabs and reducing manual intervention.


  • Full Process Automation


Thanks to the intelligent empty layer detection algorithm and auto-triggered tray retrieval commands, the system enables fully unmanned operations—from single-layer picking to completely emptying the bin—ensuring continuous production flow.


We use cookies to understand how our audience uses our site.
Transfer websites use cookies to deliver and improve the website experience, See our cookie policy for further details on how we use cookies and how to change your cookie settings Cookie policy.
Accept
Reject
Cookies settings