Welcome to the latest upgrade of our vision-guided robotics (VGR) intelligence: Bin Picking Studio 1.10 and Locator Studio 1.3.

Packed with intelligent features, enhanced performance, and integrator-focused improvements, this update delivers smarter vision, faster processing, and greater control in robotic applications.

5 Key Breakthroughs:

• AI Tote Localization Module: A dedicated AI engine for precise tote detection – even with deformed rims.
• Neural Network Switching: Configure multiple networks in one solution for optimal performance per vision system.
• Parallel MultiView Scanning: Red and blue lasers scan simultaneously to cut down acquisition time.
• Extended CAD Localization Insights: Gain deeper object-level data for performance tuning.
• Expanded Robot Integration: Seamlessly deploy solutions with newly supported robotic brands in Bin Picking Studio – HD Hyundai and Epson.

Dedicated Tote Localization

A key enhancement to the AI localization engine is the introduction of a dedicated AI module for tote localization. Building upon our impressive Depalletization solution for boxes, this module expands the automation capabilities, specifically engineered for logistics applications that involve the manipulation of standardized totes.

To enable precise recognition, picking, and relocation of these totes, the module employs a powerful neural network combined with advanced post-processing algorithms. The system robustly localizes totes in any condition: whether empty, filled with products, or even with rims bent from extensive use. 

The AI rapidly processes the scene, determining the tote’s center and optionally pinpointing its specific orientation by referencing distinctive features on the tote rim or its contents.

New tote localization capability expands our AnyPick software module in Bin Picking Studio & Locator Studio, now offering: 

• AnyPick module for Boxes (Depalletization): An AI-powered solution that localizes and depalletizes over 0,5 million types of boxes at an impeccable speed of 1300 units per hour, with up to 99.7% pick-rate. 
• AnyPick module for Totes (Depalletization): Now you can expand your logistics operations beyond just boxes, introducing a rock-solid solution for totes. 
• AnyPick module (Custom): If “out-of-the-box” solutions for boxes and totes won’t do the trick and you’ll need to flexibly deal with a wide variety of mixed goods, we offer a custom training of neural networks to match your logistics application. 

Seamless Neural Network Switching – Click to Switch! 

Our AI localization engine now provides robust support for managing multiple neural networks within a single solution. 

This allows you to fine-tune performance, especially in complex applications where distinct tasks or objects need specialized neural networks: whether pre-trained or custom-trained. 

You can now strategically assign a dedicated network to each vision system, ensuring every component operates using the AI model best suited for its specific requirements.

Not only can you deploy AI localization of boxes, totes or custom-trained neural networks, now you can switch between them on the fly, as you go! 

Parallel Multi-Perspective 3D Scanning

The CAD MultiView feature in Bin Picking Studio & Locator Studio now offers a substantial performance leap with the integration of Photoneo’s blue-laser-based 3D sensors. This enhancement empowers MultiView to execute parallel scans when users configure it with multiple static sensors utilizing distinct red and blue laser technologies. 

By triggering these scans simultaneously, this method drastically shortens the overall scan acquisition time, working seamlessly because the red and blue lasers wavelengths function independently without causing mutual interference.

By pairing two 3D sensors – one equipped with a red laser, the other with blue – you can now capture comprehensive 3D data simultaneously and completely interference-free. This breakthrough dual-laser approach fundamentally redefines efficiency for demanding tasks such as bin picking and the creation of detailed digital twins.

• Dramatically Halved Cycle Times: Achieve synchronized scanning from both devices. This parallel acquisition can effectively cut your data capture periods in half for many applications, significantly boosting throughput.
• Unparalleled 3D Data Completeness: Capture more robust 3D data by leveraging two distinct perspectives simultaneously. This is crucial for overcoming occlusions and ensuring accuracy in complex object handling, intricate assembly tasks, and high-fidelity digital replication.

Furthermore, the system is optimized automatically:

• When a combination of Blue and Red devices is used, these devices perform parallel scanning.
• When devices of the same color are used, they will scan sequentially (maintaining previous behavior).
• In a three-device scenario, such as one Red and two Blue devices, the two Blue devices will scan in sequence, and this sequential Blue scan will occur in parallel with the Red device’s scan.

Automated Diagnostics for Faster Fine-Tuning

To streamline the critical fine-tuning phase of application integration, the system now automatically saves 3D scans captured during instances of suboptimal performance. 

This opt-in targeted data collection proves invaluable, especially when initially configuring a solution. These automatically preserved scans provide integrators with the crucial diagnostic material needed to precisely analyze and adjust relevant parameters, directly enhancing the application’s overall performance and significantly accelerating the optimization process.

Enhancements in CAD Localization Engine

The CAD localization engine now furnishes users with more extensive details for each localized object, significantly enhancing diagnostic capabilities. This newly available, granular information offers deeper insights, empowering integrators to more precisely identify parameters that require adjustment. 

Consequently, users can more effectively troubleshoot and optimize their configurations, leading to substantial improvements in overall localization performance.

Performance and Configuration Improvements

This release introduces significant enhancements to both operational performance and system configuration.

• Enhanced AI Post-Processing: All post-processing algorithms within the AI localization engine see improvements, including a novel approach for computing gripping points on localized boxes.

• Maximized Sensor Connection Speed: The sensor connection process has been entirely reworked to operate at maximum possible speeds.
• Configurable Object Prioritization in Bin Picking Studio: Ensures stable sorting while still benefiting from results computed in the previous run. This boosts time-critical applications like depalletization, where consistent sorting is required.
• Fine-Tuned Hand-Eye MultiView Meshes: Users can now adjust mesh properties created by the hand-eye CAD MultiView system to optimize input quality for the CAD localization engine.
• Optional DNS Configuration for Robot Interface: Network DNS can now optionally be configured for the Robot interface, allowing advanced networking configuration.

Key GUI Improvements

The user interface across the studios receives substantial upgrades in this release, focusing on improved usability, more powerful diagnostic tools, enhanced data interaction, and greater customization options.

• Redesigned Deployment Message Console: Experience a completely overhauled message console on the Deployment page, featuring visual and functional upgrades, message filtering capabilities, and optional millisecond timestamp accuracy.
• Persistent 3D Visualizer Data: Localized objects displayed in 3D visualizers (in both the localization interface and on the deployment page) now persist, automatically reappearing after a page reload for improved workflow continuity.
• Streamlined Simulated Calibration in Bin Picking Studio: Point cloud handling during simulated calibration is significantly boosted, with the point cloud now moving in conjunction with the sensor for a more intuitive and efficient calibration process.
• Easier AI Object Selection: Selecting objects detected by the AI localization engine is now more intuitive; users can click anywhere on an object’s segment.
• Manual Scan Saving from Web Interface: Users can now manually save scans, sets of scans, or meshes (from MultiView) directly from any 3D visualizer through its Scan widget.

• Enhanced Integrator Customization: New web interface customization options empower integrators to apply brand-specific color schemes, integrate logos, and define custom error messages, streamlining client communication and troubleshooting.
• German Language Support: The web interface is now fully localized in German, enhancing accessibility for German-speaking users.

Expanded Support For Popular Robotic Brands 

This release expands robotic integration capabilities, offering users enhanced flexibility for their automation projects. Bin Picking Studio and Locator Studio now provide full support for both HD Hyundai Robotics and Epson robots.

Furthermore, Bin Picking Studio’s robot model database has been significantly expanded to include a wider array of models from both established and newly integrated robotic brands.