BPS Change Log

Bin Picking Studio 1.12

Bin Picking Studio 1.12 by Photoneo was released in July 2026.
This version includes the following new features and enhancements:

What’s New in This Release

In-App 3D Model Creation from Scans
CAD-based localization inherently requires a 3D model of the target object to match candidates in the acquired point cloud. Previously, lacking a pre-existing 3D CAD model meant a bottleneck in setting up the localization process. This release eliminates the strict dependency on external CAD files by allowing you to create a 3D model directly from a scan. Simply connect your 3D sensor, trigger an acquisition, and generate the model natively. You can then refine this model using the newly integrated Mesh Editing tools. Additionally, we have expanded our file import support to include the .PLY format.

Key Benefits:

  • Accelerated Setup: Bypasses the need to source or design external CAD files before starting.

  • Ideal for Consistent Angles: Perfect for structured or semi-structured applications where objects are consistently viewed from the same angle the model was created. (Note: For fully random object orientations, we still recommend creating the 3D model using Photoneo 3D Meshing software).

  • Broader Compatibility: Direct support for .PLY formats offers more flexibility in how you import existing models.

Integrated Mesh Editing Tools
Raw 3D models—whether imported from an external source or created directly from a scan—often contain noise, irrelevant background features, or overly complex geometry that can negatively impact localization accuracy. To address this, we have introduced a native suite of Mesh Editing tools to help you refine your localization models without leaving the software. Whether you are using a scanned model or an imported CAD file, these tools allow you to crop out noise, remove secondary objects, apply basic filtering, and adjust face normals and the origin point.

Key Benefits:

  • Higher Precision: A perfectly refined 3D model significantly improves overall localization performance and reliability.

  • Streamlined Workflow: Eliminates the need to export models to third-party software for basic geometric cleanup and adjustments.

AI-Powered Object Categorization
Traditional localization struggles in applications where the target objects vary greatly in shape and cannot be defined by standard CAD models. While neural networks can recognize these diverse objects, controllers often need more granular data than just a simple “object found” to execute complex sorting or decision-making logic. We are taking our AI-localization to the next level with Object Categorization. The on-demand retrained neural networks can now assign recognized objects into specific user-defined categories (or classes). This categorical data is then reported directly to the robot controller alongside the object’s spatial coordinates and properties.

Key Benefits:

  • Advanced Decision Making: Empowers your robotic application to execute complex sorting, routing, and handling logic based on object classes.

  • Continuous AI Enhancements: Includes under-the-hood algorithmic improvements that make the neural network faster and more robust at handling variable items.

Fast-Scan Feature in MultiView
Cycle time is one of the most critical metrics in production. Triggering a full 3D scan and point cloud evaluation during every cycle introduces unnecessary latency, especially when many objects in the bin haven’t moved since the last pick. To optimize this, the Fast-scan feature is now fully supported in MultiView configurations utilizing multiple statically mounted sensors. Fast-scan rapidly acquires a 2D texture image of the scene and compares it to the previous cycle. If an object is evaluated as unmoved across all sensor viewpoints, it is immediately accepted as a valid picking candidate without requiring a full 3D scan computation.

Key Benefits:

  • Minimized Cycle Times: Drastically reduces processing time by skipping unnecessary 3D acquisitions for static objects.

  • Compounding Efficiency: When combined with the parallel scanning capabilities of Blue & Red sensors introduced in previous releases, Fast-scan pushes your application’s cycle time to the absolute minimum.

System Optimization & Usability Enhancements

Driver & Backend Optimizations

  • Upgraded to PhoXi Control 1.17.2: This release integrates the latest PhoXi Control driver capabilities

    • Refreshed UI of Settings Assistant: It now prioritizes profiles by scanning time, instantly displays key data, recommends a “suggested” profile, and fully supports new PhoXi 3D Scanner Gen3 sensor parameters, which ultimately reduces setup time.

    • Maintenance tool update: Using multiple identical marker patterns simultaneously is now allowed, drastically speeding up the maintenance process. The required number of scans is reduced from 10-15 to 3-4 in most cases, making sensor maintenance 3-4x faster for customers.

    • Visual 3D ROI: We extend the legacy numeric input fields with a user-friendly, visually adjustable 3D bounding box. This makes defining the Field of View (FoV) highly intuitive, fast, and error-free for users of all experience levels.

  • AI Algorithm & Gripping Point Optimization: Computation time for gripping point placement has been significantly reduced for selected placement methods, accelerating the cycle time of AI-driven bin picking tasks.

  • Improved Dynamic Mode in Hand-Eye MultiView: Real-time object tracking in dynamic environments has been upgraded. The updated tracking algorithms are significantly more robust, ensuring reliable candidate persistence even in rapid or challenging handling scenarios.

  • Optimized Scanning Profile Switching: Scanning profiles are now uploaded to the sensor using their original names (without the suffix referencing the specific vision system ID). As a result, when multiple vision systems share the same profile, only a single profile is uploaded to the physical sensor. This eliminates the need for physical profile switching on the device when transitioning between vision systems.

Robotic Integration & Support

  • New Robot Brand Integration - Estun: We are excited to announce our official support for Estun robots. A dedicated Estun robot module is currently under development to provide a native plug-and-play experience.

  • Expanded Robot Model Database: The built-in robot model library has been updated with new, highly detailed models from various existing robotic partners to simplify workspace planning and collision avoidance setup.

Enhanced Default Scanning Profiles

To simplify your initial setup, we have tuned our default scanning configurations for peak performance in specific environments:

  • Calibration Profiles: Fine-tuned to perform robustly across a much wider range of ambient lighting conditions, preventing calibration failures due to glare or low-contrast environments.

  • AI Application Profiles: Pre-configured with optimal color reprojections to provide highly accurate, balanced image inputs specifically suited for neural network training and inference.

GUI & Visualizer Enhancements

  • Global Settings Override Indicator: Vision systems that override global configurations are now instantly recognizable via a visual indicator, allowing you to identify custom parameters at a glance.

  • Reworked AI-Localization Object Rendering:

    • To improve visual consistency, localized objects are now rendered using identical, matching colors in both the 3D point cloud visualizer and the 2D texture visualizer.

    • Objects rendered in the 2D texture visualizer are now fully clickable, allowing for rapid inspection and detailed data lookup.

  • Advanced Point Cloud Coloring Options:

    • Added a broader selection of solid color themes.

    • Enhanced the height-ramp coloring mode. Users can now manually configure the range of the height ramp and choose to either saturate, dim, or entirely hide data points that fall out of the defined range.

  • Back-Face Mesh Rendering: The back-faces of CAD meshes are now rendered in black (previously, they were invisible). This provides immediate visual feedback on normal orientations, which is critical for configuring the CAD-based localization engine correctly.

  • Incremental Input in 3D Visualizers: Precision object manipulation is now easier. Wherever 3D objects are rotated or translated, you can simply click on the draggable 3D marker/gimbal to input a specific numerical increment to be added directly to the current value.

  • General Polish: Includes various minor GUI enhancements, stability improvements, and bug fixes to ensure a smoother development workflow.

Bin Picking Studio 1.11

Bin Picking Studio 1.11 by Photoneo was released in October of 2025.
This version includes the following new features and enhancements:
PhoXi Scanner Gen3 support & PhoXi Control 1.16
The integrated PhoXi Control version has been updated from 1.14 to 1.16. This is the first PhoXi Control & firmware version built to power our next-generation scanner. The Gen3 is packed with cutting-edge technology designed to deliver unprecedented performance:
  • Higher Resolution: A powerful 5.1 Mpx sensor delivers 60% more 3D points for incredibly detailed scans.

  • Any Lighting Conditions: Thanks to new Blue Laser technology and HDR support, the Gen3 can perform flawlessly even in direct sunlight.

  • Integrated RGB Camera: Capture rich, detailed scenes with up to 8 Mpx color images.

  • Faster & More Robust Scanning: Exposure times are cut by up to 10x, and a 20 dB higher dynamic range ensures robust performance.

  • Extended Scanning Range: Now you’re able to scan up to 6 m distance.

Additionally, the newest PhoXi Control brings an updated Maintenance tool and various user experience improvements.

Multiple localization bounding boxes
The latest update to the localization engine introduces the capability to configure multiple bounding boxes within a single localization profile.
This enhancement significantly reduces the solution’s complexity (fewer vision systems required) and speeds up configuration time.
The active bounding box is selected by the robot on the fly during solution runtime via a new action request implemented in the latest robot module version.
Fine-alignment of scans in Multi-sensor calibration
Calibrating static sensors in a CAD MultiView configuration can be challenging, especially with larger sensor models.
To overcome the inaccuracies often introduced during the standard calibration process, we’ve implemented a new feature: Calibration Fine-Alignment.
When applied, the feature uses the entire captured scene to fine-tune the alignment of individual sensor scans, significantly improving the overall calibration accuracy.
CAD localization profile optimization
The size of a CAD localization profile can become quite large, as it stores a pre-trained model for the specific CAD geometry, with the final size depending on the configured parameters.
With newly incorporated optimizations within the CAD-based localization engine, the profile size can be decreased to as little as 10% of its original size.
Localization profiles originating from older versions are automatically migrated and optimized.
Software modules page
To provide a comprehensive overview of available functionalities, the web interface now includes a new dedicated page for Software Modules.
This page displays the full range of expansion modules available for your Studio version, along with their current unlock status.
To unlock an expansion software module, please contact your sales representative.
Robotic brands and robot models
We are excited to announce basic support for Nachi and Techman robots! Dedicated robot modules for these brands are currently under development.
Additionally, the robot model database has been expanded with new models from both these new brands and our already supported robotic partners.

Finally, alongside the new action request for localization bounding box selection, we have implemented the Communication Check action request. This new request allows the robot program to verify the connection status to the vision controller, ensuring communication is working correctly before proceeding.

Documentation Center

Documentation Center by Photoneo was released in late June of 2024 with the following new features and enhancements:

New features

MultiView Module for CAD Localization
The MultiView module enhances Bin Picking Studio by enabling the capture of scene scans from various viewpoints. These scans are then merged to create a comprehensive point cloud for improved localization.

MultiView configurations:

  • Static MultiView Employs statically mounted 3D sensors. One primary sensor is calibrated with the robot, while secondary sensors are calibrated relative to the primary. Individual scans are merged with high precision, resulting in a detailed scene point cloud.

  • Hand-eye MultiView Utilizes the standard hand-eye approach with a robotic arm holding a Photoneo 3D sensor. Instead of a single scan, the MultiView module captures the scene from numerous viewpoints to create a complete point cloud.

Applications - the MultiView module addresses limitations of the single-view approach:

- Thin or highly reflective objects difficult to scan from a single perspective
- Requirement of specific model features for precise localization, obscured in a single view
- Self-occluding objects
- Overly large objects requiring multiple scans
- Need for higher detail quality, achieved by using a sensor with lower point-to-point distance and scanning volume and capturing the scene multiple times
The MultiView module for CAD localization has to be ordered separately as an extension module.
Feature Alignment
The CAD localization engine receives a boost with feature alignment functionality. This function refines object orientation based on the alignment of its key features. Users highlight key features on the CAD model, allowing the localization engine to prioritize these areas. Additionally, feature axes of symmetry are defined to guide the engine in searching for object invariance with the highest overlap in the key feature region.
Calibration Sharing
Vision systems often utilize the same 3D sensor and calibration. To avoid individual calibration for each system, a new global setting, “Calibration Sharing,” has been implemented. When enabled, all systems using the same 3D sensor and calibration configuration will have their calibration parameters updated together (calibration matrix, profiles, automatic calibration settings).
Reusing Scans
Previously, each vision system with the same sensor required a unique scan for localization, even if the scene was identical. Now, the last scan from each sensor is stored in memory and available for use by other vision systems employing the same sensor.

User Experience Improvements

  • Simulated calibration Allows virtual adjustment of the 3D sensor pose when a robot is unavailable for real calibration. This enables the testing of solution configurations (path planning) with recorded 3D sensor scans.

  • Up to 4 sensors directly connected to the vision controller Redesigned Network page allows configuration of all available Ethernet ports for robot, local network, and multiple sensor connections.

  • Extended Fixed plane offset functionality Fixed plane can now be defined perpendicular to any base robot axis (X, Y, Z).

  • Studio navigation bar and solution wizard now have a fixed floating position.

  • New Picking priority method: Gripping point priority

  • Bounding box visualization around AI-localized objects.

  • Change scene state request added to Deployment action request toolbar (Simulation mode).

  • Redesigned user interfaces for Localization, Network, and Vision System pages.

  • General GUI improvements and bug fixes.

System Improvements

  • PhoXi Control updated to version 1.13.1.

  • Host OS - Ubuntu - upgraded to version 22.04.4 with:

    • Up-to-date security patches

    • Latest versions of supporting third-party applications

Robotic brands, Communication protocol and Robot modules

Expanded Robotic Brand Support
We’re excited to welcome back Mitsubishi to our family of supported robotic brands!  Additionally, RobCo robots join the fold as our newest member.
Enhanced Robot Model Database
The robot model database has been expanded to include a wider range of robots.
Communication Protocol v1.9
Communication protocol v1.9 brings support for new robot action requests:
  • Trigger scan: Allows adding a new scan within the Hand-eye MultiView configuration.

  • Localize on the last scan: Enables running localization on the most recent sensor scan.

Robot Module Updates
Updated robot modules for supported robotic brands based on the newest version of the communication protocol will be available shortly.
All robot modules based on the previous versions of the communication protocol remain supported.

Bin Picking Studio 1.8.1

The first patch version of the Bin Picking Studio 1.8 was released in early October 2023 and includes bug-fixes for following issues:

  • Broken automatic Deployment page reload upon solution stop/start via robot action request after several minutes of robot inactivity

  • Input of borderline values for rotation of a collision object is not possible

  • Numeric values in the deployment Inspector and the Robot modules are rounded differently

  • Automatic calibration is disabled upon any change of scanning profiles

The patch also brings several usability improvements:

  • Reworked layout and enhanced usability of the localization interface

  • Added option to choose radians or degrees as the unit of angle/rotation values in the deployment Inspector

  • Solution that was deployed as last is marked with a new flag last run

  • Added option to specify a gateway on the Robot interface

  • Added automatic solution backup option - the solution is exported every time it is deployed

  • PhoXi Control was bumped to the latest version 1.11.1

Bin Picking Studio 1.8.0

Bin Picking Studio 1.8.0 by Photoneo was released in early July of 2023 with the following new features and enhancements:

New features

AnyPick module for Boxes - solution for Depalletization
AnyPick module for Boxes extends the functionality of Bin Picking Studio or Locator Studio with the ability to recognize unknown boxes based on pre-trained neural network. This functionality allows Studio to be deployed also for Depalletization applications.

AnyPick module for Boxes has to be ordered separately as an extension module. For more information and inquiries, visit https://www.photoneo.com/depalletization.

When Bin Picking Studio is extended with the AnyPick module, user can create solution of two types:

  • based on CAD-object recognition

  • based on AI-object recognition (boxes)

The configuration wizard in AI-based solution guides the user only through relevant configuration steps. The gripping point is chosen automatically based on the visible surface of the box. The output data provides information on dimension and orientation of the box to allow precise placing.

AnyPick module for Boxes is supported (trained) in combination with PhoXi 3D Scanners.

Automatic robot-camera calibration
A calibration routine verified by the operator can now be repeated automatically using the new action requests in the robotic API:
  • Start the automatic calibration

  • Save the result of the automatic calibration

  • Stop the automatic calibration

The semi-automatic calibration introduced in version 1.6 can still be used to safely verify the calibration routine.

Reporting of vision system status to the controller

Added request to the robotic API - Get vision system status - to provide information about:

  • number of localized objects

  • number of ready objects

  • state of the processing pipeline (in progress/finished)

This information is especially useful in combination with multiple vision systems to prioritize picking from the most full bin or to receive early information about almost empty bin.

User experience and system improvements

  • Redesign of Vision system page into into three clear views:

    • Configuration

      • setup of the 3D sensor and localization scanning settings

      • localization and localization settings

      • calibration type and calibration scanning settings

    • Calibration overview

      • information about current robot-camera calibration

      • improved manual edit of the calibration matrix

      • option to set-up automatic calibration

    • Settings - optional settings overriding global defaults

  • Update of Deployment Inspector - showing object dimensions (for AI based localization of boxes) and gripping point priority (for CAD based localization)

  • Added support for 14 new robots (total 217 supported robot models)

  • CAD localization improvements

    • Added parameter for special fine-tuning of CAD localization

    • Added option to specify when the object is considered occluded

  • General GUI improvements and bugfixes

  • PhoXi Control update to version 1.10, support for all Photoneo 3D Sensors

  • 3rd party applications updates:

    • Google Chrome 114

    • Meshlab 2021.05

    • CloudCompare 2.11.1

User experience and system improvements

Communication protocol v1.8 adds support for the new action requests in the robotic API for performing Automatic robot-camera calibration and reporting vision system status.

Updated robot modules based on this version of communication protocol are available for:

  • ABB

  • Doosan

  • Fanuc

  • Kawasaki

  • KUKA KRC

  • Universal Robots

  • Yaskawa

All robot modules based on the previous versions of the Communication protocol remain supported.

Bin Picking Studio 1.7.1

The start of December 2022 brings the first patch of the Bin Picking Studio 1.7.

We improved system stability and fixed several issues:

  • Error 500 when BPS was turned off abruptly

  • Fast scan parameter Changed pixels fraction threshold scaled by factor x100

  • Initialization request not stopping localization for vision systems other than the one for which the request was called

  • Calibration page not updated when the robot adds a calibration point

  • Objects not correctly displayed in the hierarchy in Visibility and Axis widgets of 3D visualizers

  • Configuration of a vision system not saved correctly upon first save action

We also made some improvements:

  • Changes to the Robot interface are applied faster

  • The Inspector on the Deployment page was redesigned

  • The deployment shortcut in the solution wizard has been replaced by a button

Bin Picking Studio 1.7.0

Photoneo releases Bin Picking Studio 1.7.0 - the new version of the company’s bin picking tool powered by 3D vision and robotic intelligence. The upgrade to 1.7.0 extends bin picking possibilities and improves user experience through the following new features and enhancements:

System improvements and changes

  • When set to the DHCP client option, the network interface displays the IP address assigned by the DHCP server

  • When configuring a static IP address on the network interface, it is also possible to define DNS servers

  • BPS can be simultaneously operated locally (Vision controller) as well as remotely (remote computer) through the BPS web server. It can be used with a static or dynamic IP address on the Network interface.

GUI Improvements and changes

  • Communication interfaces are now located on a single page

  • Option to visualize the current state of the robot on the Environment page

  • Improved layout of the PhoXi interface

  • The ready-to-deploy solutions on the Deployment page are now displayed in a clear table instead of a drop-down list

  • Collision objects in the Visibility and Axis widgets of the 3D visualizers are now hierarchically displayed as defined on the Environment page

New features

MotionCam-3D integration
Bin Picking Studio 1.7.0 supports the integration of MotionCam-3D!
The camera can be used in both the scanner mode and the camera mode. Each mode provides its advantages, depending on concrete application requirements.
Scanning profiles
Objects made of different materials or having different surface finish and shape often require different scanning parameters. Now it is possible to define a unique scanning profile - a set of scanning parameters - for each vision system.
This scanning profile is automatically applied during scan acquisition. This is especially useful when a single scanner is used for scanning different objects. The scanning profile support is available for both the PhoXi 3D Scanner and MotionCam-3D.
Complete vision system settings
The vision system is a largely independent unit that often requires unique values for different settings. Now it is possible to define a custom value for almost any setting of every vision system.
Multiple scene states
A change in the state of the robotic cell during bin picking can now be reflected in the bin picking solution by defining multiple scene states.
When using multiple scene states, it is possible to assign an object to a specific scene state so that it is present only when that scene state is active.
Switching off scene states is requested by the robot and can be done through the new Change scene state request introduced in the Communication protocol v1.7.0.
Calibration without the Robot State Server
Calibration of a vision system requires a functional Robot State Server that provides the state data of the robot. If the Robot State Server is not available for some reason, there is an option to input the robot state manually by rewriting the TCP data from the teach pendant of the robotic controller in the BPS web interface.
This feature is currently available for ABB, Fanuc, Kawasaki, KUKA KRC, and Universal Robots.
New localization stop conditions
Localization searches for objects until a stop condition is met. In previous BPS versions, the only stop condition was the overall timeout. Now users can stop localization after a specific number of objects has been localized or when no new object has been localized for a certain time period.

Communication protocol and robot modules

We introduce a new version of the Communication protocol - v1.7.0, which brings a new robot request: Change scene state. New robot modules based on this version of communication protocol are available for:

  • ABB

  • Fanuc

  • Kawasaki

  • KUKA KRC

  • Universal Robots

All robot modules based on the Communication protocol v1.6.0 remain supported. We have also added Doosan among supported robotic brands with the robot module based on the communication protocol v1.6.0.

Documentation extensions

The documentation menu has been reworked.
There are new documentation pages that cover the support of 3D sensors and Scanning profiles and the mechanism of usage of bin picking solution settings.
Many pages have been updated - the Environment page has been extended with the new Multiple scene states section and the Calibration page has been significantly improved.

Bin Picking Studio 1.6.1

We are happy to announce the integration of Stäubli among the supported robotic brands.
Besides that, we fixed a few issues that managed to slip past our radar.

Bin Picking Studio 1.6.0

At the end of 2021, Photoneo released the Bin Picking Studio 1.6.0. This release brings major system changes, several new features, and further improvements to the overall user experience:

System improvements and changes

  • BPS is now running on Ubuntu 18.04 and the latest version of Google Chrome and TeamViewer.

  • BPS GUI is running smoother and faster. Execution times of several actions have decreased significantly; deployment start-up time of a standard bin picking solution has been reduced under 10 seconds and the deployment stop time takes just a second. That makes the change of the deployed solution from the robot controller several times faster than in the previous versions of BPS.

  • Connection through a web browser from another computer via a local network is fully supported.

  • Use without a connected monitor is fully supported (web browser and VNC access from another computer via a local network, TeamViewer).

  • Improved network settings management which prevents conflicting settings between the BPS and the host system.

  • Autostart feature has been replaced with automatic last state recovery.

GUI Improvements and changes

We are always making some tweaks to improve the user experience. The major ones include:

  • Duplication of an existing vision system.

  • Visualization of the gripper’s volumetric model in the Inspector.

  • Visualization of the localization bounding box on the Deployment page.

  • Visualization of colliding parts of the robot and parts of the environment in the Inspector.

  • Duplication of an existing environment object and replacement of the CAD model of an existing environment object.

Other minor tweaks:

  • Bulk removal of solutions.

  • Floating save button on the Settings page.

  • Displaying of solution/vision system ID next to its name for easier identification.

  • Changes on the Settings page are preserved when they cannot be applied during saving because the solution is currently running in the deployment.

  • Inspector widget color scheme changed for better readability.

Some pages have changed or improved layout:

  • The Picked object page now uses the same layout as the Gripper page.

  • Reworked and improved vision system page layout.

  • New Help panel layout.

  • Robot interface and Robot controller tabs on Network page joined into single Robot tab.

New features

Vision controller - Robot communication
Communication between the Vision Controller and the robot controller is now running persistently. The robot controller is able to connect to the BPS at any time and its brand is detected automatically. While connected, the robot controller is able to call any of the supported requests. Following new requests are supported in BPS 1.6.0:
  • Add calibration point

  • Start solution

  • Stop solution

  • Get running solution ID

  • Get the list of available solutions

Support for the robotic brands in this release is limited to ABB, Fanuc, Kawasaki, KUKA, Universal Robots and Yaskawa. BPS now also supports KUKA LBR iiwa.
The robot model database has been further expanded with many new robot models.
Multiple grasping methods
BPS now enables the definition of a specific grasping method for each vision system. This enables the use of different picking strategies for each vision system which might be needed due to differences in bin shape, picked objects, or other factors.
Semi-automatic calibration
After starting the calibration procedure in the BPS, the robot is able to call the “Add calibration point” request. The user is able to write a simple robotic program to automate the calibration process. After the last point has been added, the user just verifies and saves the result in BPS.
The robot module for each brand comes with an exemplar calibration program.
Robot workspace visualization
Environment page visualizer offers the option to visualize the workspace of the robot based on its current joint limits. This information can be used to improve the definition of joint limits to achieve the maximal performance of path planning.
Simulation mode
Bin picking solutions are now deployable in the Simulation mode. In this mode of deployment, the user is able to call basic bin picking requests directly from the Deployment page. This enables simple and fast evaluation and tuning of bin picking performance without the need to have the real robot connected.
Cycle timer
The Deployment page of BPS now offers to monitor basic bin picking statistics. The user is able to observe the cycle time based on incoming trajectory requests or scan requests as well as the average cycle time.

Documentation extensions

Documentation has been heavily reworked, pages were reorganized into a new hierarchical structure and several new pages have been added including:

  • Network

  • Robotic API

  • Error codes and error handling

  • Deployment page

Bin Picking Studio 1.5.2

The second patch release of Bin Picking Studio version 1.5 introduces Doosan Robotics into the family of supported robotic brands.
We have added 10 new robots to the robot model database and made some improvements and fixes to the documentation.

Bin Picking Studio 1.5.1

This is the first patch release of Bin Picking Studio version 1.5. It fixes several issues that the users may have encountered in version 1.5.0 and also introduces some handy new features:

  • Bin Picking Studio now offers a Japanese as well as a Chinese language version of the user interface.

  • The planned trajectory for an object that is currently being picked is automatically visualized after a trajectory request has been made.

  • The table of defined gripping points can now be sorted for convenience.

  • The status of the connection to the Robot State Server is now visible during the teaching of a gripping point.

  • The Localization SDK has been upgraded to version 1.3.4.

  • The robot database has been further expanded by 7 new robot models.

Bin Picking Studio 1.5.0

In late October 2020, Photoneo released the Bin Picking Studio 1.5.0. Our main effort and focus has gone into improving the user experience. Following new features have been implemented:

Speed up

Thanks to the improvements in the localization engine, we reduced the average time of localization by approximately 500 ms. In addition, the path planning performance was increased to save an additional 200 ms minimum per cycle.

PhoXi Control

The new release of Bin Picking Studio comes with a new PhoXi Control 1.2.26. Besides network stability improvements, users can look forward to new, handy features useful in bin picking applications (e.g. new interreflections filter). See the list of changes in the complete release notes.

Localization

Together with the new version of Bin Picking Studio, we also release a new version of Localisation SDK numbered 1.3.3. Besides general stabilization and speed improvements, users can look forward to the following features:

  • Empty scene threshold - a new parameter that enables users to define a threshold value for a number of 3D points from the PhoXi scanner. If there are fewer 3D points available, the robot receives this information, which means that the scene is empty. In other words, there are no more objects to be picked.

  • Fast scan settings are applied also in the localization GUI during the configuration of the localization profile.

Remote access

The Vision controller is now running a VNC server for simple and effective remote control. Information about the connection to the server is summarized on a new documentation page.

Robot modules

As with each release of BPS, we extended the robot database with new robot models. We added 13 more robots to choose from.
Besides that we:
  • Reworked the KUKA module to increase the speed of communication with the Vision controller,

  • Extended the support of UR controllers; up to version 5.9.x.x for e-Series and up to version 3.14.x.x for CB-series,

  • Added the official support of ABB OmniCore controllers.

Last but not least - we announce the official support of Denso robots.

Convenience features

Besides all the features described above, we made a few smart enhancements across the entire system. These include:

  • Improved the calibration process in order to perform calibration without any difficulties,

  • Implemented automatic detection of incorrect TCP reported by the robotic controller during calibration to prevent incorrect results,

  • Tool point definition is now part of each gripper in the gripper database,

  • Object gripping points definition was moved to a separate page,

  • Environment page and Visualization page are now one single Environment page,

  • Collision checking (collision highlighting) is now available also in the joint mode,

  • Ability to download the STL of previously uploaded STL objects in the Environment builder

  • Each gripping point can have its own thumbnail displayed on the gripping point list to make the orientation easier,

  • Numerous GUI improvements,

  • Visualizers on the localization page and deployment include a new widget providing information about the last scan properties.

Settings

We enhanced the settings page and improved the UX.
There are also new, handy parameters to make your application more efficient. The first two allow you to define the behavior of a vision system after its initialization by a request from the robot:
  • Clear memory on initialization,

  • Force full scan on initialization.

The next two parameters are mainly useful in time-critical applications where the collision-checking performance can be tuned to exchange high precision for extreme speed:

  • Resolution of the point cloud volumetric model,

  • The minimum required size of a volumetric model segment.

We recognize the need for different settings for each vision system as they can often have special requirements. In this release of Bin Picking Studio, we moved the most crucial parameters from the solution settings to be unique for each vision system:

  • Gripper collision sensitivity,

  • Fast scan settings.

Furthermore, the fast scan feature has been added with new parameters that allow the user to configure the object localization from a texture image.

Documentation extensions

Documentation updates include:

  • A detailed description of the bin picking routine speed customization in bin picking tutorial for each robotic brand,

  • New page summarizing the base frame offsets for every supported robot model that requires it,

  • Page about the connection to VNC server,

  • Tutorial on how to set up the robot in various orientations (upside down, on the sidewall),

  • Tutorial for a correct tool point and gripping point setup.

Bin Picking Studio 1.4.0

In March 2020, Photoneo released the Bin Picking Studio 1.4.0. We have implemented the following new features:

Collision checking

Collision checking algorithms were completely rewritten and enhanced with the following new abilities:

  • BPS now computes collisions with the grasped object.

  • Collision checking with other objects localized in the bin – localization can detect objects even if they are not entirely visible in the scene. The entire model is rendered into the scene and used as a collision object.

  • Gripper collision sensitivity – when picking with a two-finger gripper, the robot sometimes collides with the neighboring objects. As a solution to this problem, users are now allowed to define gripper sensitivity and set what percentage of the gripper volume may get in a collision with the scanned point cloud or with any currently detected objects.

Speed-up

Speed is a high priority for every customer. We have made the following speed optimizations:

  • The waypoint computation is 100-times faster for most of our robots thanks to the brand new inverse kinematics calculation algorithm.

  • Connections to cameras are faster.

  • The solution deployment time was reduced by 50%.

Localization

Together with the new version of Bin Picking Studio, we are also releasing a new version of Localisation SDK numbered 1.3.0. Besides general stabilization and speed improvements, the users can look forward to the following features:

  • The Localization GUI now allows the definition of a bounding box where the detection will take place, leading to higher speed and increased safety.

  • A segmented image is now available, which is crucial for proper configuration of the localization settings.

Scanner on the robot

The possibility to mount a scanner on the robot, a so-called hand-eye approach, was frequently requested by our partners and customers. This allows the user to use a smaller scanner for bigger bins and thus get a higher resolution and better details.
Another significant benefit is provided by the option of variable viewpoints, which comes very handy in case one needs to take a closer look at the corners of the bin.
Attach the scanner to the end of the robotic arm or gripper, calibrate them with a brand new marker pattern calibration and enjoy the benefits of scanning from variable viewpoints.

Inspector

We have pushed “pure” debugging to another level. Besides marking the localized objects with color, we now also display complete trajectories in 3D space. The user can click on the detected objects and inspect:

  • computed grasping positions,

  • picking path stages,

  • trajectories.

You can literally replay the robot’s planned motion. Position by position.

Environment

  • Added support for drawing simple collision objects directly in the GUI.

  • Added support for resizing and scaling environment objects.

  • The robot is now displayed in the environment configuration GUI.

  • Connected and calibrated scanners are now automatically displayed in the environment.

  • Connected vision systems can be triggered and the point cloud will be displayed, enabling a quick verification of the position of objects in the scanner’s view.

  • It has never been easier to set up the scanners’ field of view thanks to the ability to display their scanning volumes.

Calibration

The process of alignment of the scanning volume with the robot’s working volume is crucial for successful bin picking. We will continuously work on improving the calibration pipeline. Besides higher user-friendliness, users can look forward to the following features:

  • Added pre-configured calibration ball, with its diameter being remembered by the system

  • Added the ability to visually verify calibration results in the embedded visualizer.

  • Added support for hand-eye calibration.

Robot motion description

Fluent movement of the robotic arm is a basic requirement of any customer. Therefore we have made a few tweaks to the path stage configuration:

  • The approach and de-approach waypoints are now mandatory.

  • Motions planned between approach and de-approach path stages are planned as linear motions.

  • The maximum number of waypoints is 11.

  • The sampling step in linear trajectories can be configured in User settings.

Visualization

In the previous version, only joint jogging was available in the visualization. Users can now jog in reference to either the coordinate system of the robot or the coordinate system of the tool. All movements respect the kinematics of the robot according to the datasheet, taking joint limits into consideration too.

Robot modules

Robot modules were also updated to improve error handling and connection reliability. In addition, new, more complete examples were added for every brand.

Convenience features

Besides all the features described above, we have made a few smart enhancements across the entire system. These include:

  • Solution attachments to keep all data in one place. Users can attach additional files, such as robot program backups or additional CAD files, into their solutions. The files will be embedded directly into PBCF solution exports.

  • The ability to choose between tiled view or table view of robots, grippers, and picked objects.

  • New user roles and a user access management section to assign different access rights to different users of the BPS.

  • Welcome page for the initial, quick setup of:

    • the installation timezone,

    • customer’s accounts.

  • If you have a picking preference, you can now let the system know by setting your Favorite gripping point - by assigning it a high, medium or low status.

  • Troubleshooting of network connections is now easier due to new, descriptive error handling.

Settings

Set new, handy parameters to make your software more efficient:

  • Inhibit the robot from picking partially occluded objects which often causes blocking.

  • Specify the direction in which the objects are to be picked. You can now configure the picking priority vector in robot space.

  • Disable or enable the possibility of a collision with the picked object, depending on the way the object is handled.

Documentation extensions

For easier understanding, the documentation is full of pictures and descriptions supporting the text. It has been extended with new sections, mainly related to path planning, the Environment page, the calibration process, as well as user roles and permissions.

Bin Picking Studio 1.3.0

In July 2019, Photoneo released the Bin Picking Studio 1.3.0. Our main effort and focus has gone into improving the user experience. We have implemented the following new features:

  • From now on the user will be able to import solutions from versions 1.0.0, 1.1.0 and 1.2.0 and thus save a lot of time thanks to the import converter which enables backward compatibility.

  • Path Planning is not a simple task and we are continuously testing and evaluating new approaches. While one approach suits certain types of applications, it might not be the best solution for others. From now on the customer can decide for and select one of the following approaches:

    • Standard - the first option provides high speed, however, it requires a precise definition of robot joint limits for optimal trajectory planning.

    • Advanced - in the second option, the path planner is able to compute much more fluent trajectories, however, the overall speed of path planning is slower.

  • Sometimes the environment might not be as difficult as to justify complex path planning. We have integrated a path planner selection in 1.3.0 where the user can choose from the following options to influence the cycle time:

    • Path planning for all trajectories – this configuration has been taken over from the previous versions of the Bin Picking Studio. The trajectories between all waypoints are computed by the path planner. For all of them, collision checking is enabled. This is the safest way to execute bin picking routines and BPS studio users are strongly encouraged to use this setting.

    • Path planning for linear trajectories – All BPS users are recommended to use linear trajectories from the approach waypoint to the grasp pose and back to the de-approach waypoint. With this setting, all path plans will be omitted with the exception of these 2 linear trajectories.

    • Path planning off – This configuration poses the greatest risk. This setting is not recommended for end users unless they are very advanced

  • We have integrated a solution switch command to handle different bin picking tasks on different working shifts. It is now possible to switch a running BPS Solution over to another one in the robotic controller with the solution switch request. Please refer to our extensive robot module documentation for more details on the implementation of your robot brand.

  • Over the past releases, the community of experienced Bin Picking Studio integrators was formed. These users can now make use of the following, more advanced settings:

    • Gripper collision sensitivity – if your gripper collides with the scanned point cloud, lower the number. If too many of the detected parts are rejected because of gripper collisions, raise this number.

    • Use a fast scan – a fast scan dramatically lowers the scanning time by triggering only the fast acquisition and computing only those parts of the scene which have been changed in comparison to the previous scan.

    • The above-mentioned setting also contains a configuration option in the form of a threshold number of the parts. If the number of the parts in the previous run is lower than this threshold, the system will trigger a full scan anyway.

    • The full scan frequency defines the proportion of fast scan acquisitions as compared to full scan acquisitions.

    • Trajectory request timeout – the maximum time allowed for the path plan computation.

    • Translation threshold – the minimum part displacement allowed to identify a part translation from the current scan as a duplicate of a part from the previous scan.

    • Rotation threshold – is the same as the offset threshold, but related to the rotation of a part.

    • Tool point vs. robot base Z-axis angle threshold - this setting defines the maximal grasp angle.

  • PhoXi interface improvements – the BPS now displays scanner statuses and disables the connection to scanners with incompatible scanner firmware versions.

  • Gripping point IDs – users are now able to redefine the IDs of gripping points in the grasping method.

  • All numbers reported in calibration and robot joints are now rounded.

  • Activity logging in the solution - from now on, users can inspect all changes in the solution settings.

  • User management – authorization to create multiple new users accounts within the organization unit.

  • Deployment page console – multiple new messages have been added to simplify the solution debugging.

  • Photoneo engineers are ready to help you with the deployment process. In case of error, press the Download logs button on the deployment page and send the file to our support department for a quick analysis.

  • We have enhanced the web 3D visualizer to offer better investigation options, camera controls and significantly better rendering performance. The number of point cloud points in the visualizer has been increased by ten times.

  • All help texts and tooltips have been carefully reviewed.

  • The documentation has been updated with new sections which contain detailed descriptions of changes in previous versions and picking path stages configurations.

  • The bin-picking training material is now available as built-in documentation.

  • Virtual robot – when an object cannot be picked, the user can not only see the status of the part and the reason why computation has failed but, in some error cases, also the virtual robot in a problematic state. Debugging has never been easier!

  • Another major feature is called Teaching Gripping point. Setting up a gripping point is even easier.

  • Last but not least the family of robotic brands has been extended by beta support of selected Yaskawa controllers and Nachi controller type CFD.

Bin Picking Studio 1.2.0

In March 2019, Photoneo released the Bin Picking Studio 1.2.0. Based on the feedback from our customers, our main efforts were focused on improving user experience. However, improvements to stability and usability were also on the radar of our development team. The result are 68 bug fixes. You can look forward to the following new features:

  • Because a simple setup of one robot, one scanner and one part was not a challenge anymore, we decided to introduce multiple vision systems and multiple picking objects. The user can configure up to 4 objects in a single solution, whereby each object is assigned to a dedicated bin and PhoXi scanner.

  • We have launched the PhoXi Control integration. Bin Picking Studio offers web-based scanner user interface which provides basic functionalities such as device list, software trigger, point cloud, depth map or texture preview.

  • We have also integrated the Localization configurator. Bin Picking Studio now includes a web-based user interface for configuration and verification of Photoneo’s Localization parameters. All user settings are available here.

  • For user convenience, robot joint angles can now be viewed also in degrees, not only radians.

  • Previously there were issues with setting proper TCP/IP ports on the network page. From now on, the correct port configuration for the robot is determined automatically based on the robotic brand. Just use the default parameters provided by our example programs and you will be ready to go.

  • We have added German language to the studio. More languages will come with the next releases!

  • Every system requires a restart from time to time. This is not an issue anymore thanks to the new Autostart feature. It is possible to run the Bin Picking Studio right at the Vision controller boot with one pre-selected solution.

  • The family of supported robotic brands has been joined by Mitsubishi.

  • There are 12 new robotic manipulators available, including new UR e-series.

  • Do not miss our new 3D visualization on the deployment page. Besides the Robot with its environment and point cloud representing the scene, the localized parts are present, including a color code and related statutes. Debugging was never easier.

  • The path stages have been extended by new waypoints which makes the configuration of the extraction procedure more flexible for the user. We have introduced fixed waypoints in a layer defined by the user. There are also further new waypoints with a relative offset to Start and End position.

  • Finally, the documentation is easily available, including a built-in robot documentation. It has been updated to be more precise and descriptive.

Bin Picking Studio 1.1.0

Following the new releases of Localization SDK 1.2.2 and PhoXi Control 1.2.6, we announce a second release of our Bin Picking Studio.
In this version we focused mainly on improving stability and on implementing feedback from our pilot users.
  • Among other features, the main improvements include:

  • New interfaces implemented in KUKA, Kawasaki and Fanuc manipulators

  • New robot models

  • Enhanced visualizer

  • Multiple major and minor bug fixes and improvements to stability and speed

  • Gripping point deactivation

  • Verification of a collision with the scan - gripper cloud

  • Support of multiple vision systems

  • Advanced collision objects / environment management

  • Built-in robot-setup documentation

  • Addition of more descriptive help texts

  • PhoXi 1.2.6 interface

  • Built-in configuration of the network interface

  • Robot-page redesign

We appreciate the support and feedback from the testing group of our first pilot users.

Initial Bin Picking Studio release 1.0.0

First release of Photoneo’s Bin Picking Studio for internal purposes.
Our vision and aim is to transform our experience and knowledge gained from projects delivered directly by us into a product which enables our partners to deploy robust and reliable Bin Picking applications.
In this first version we provide 50+ robotic models from vendors such as ABB, UR, Staubli, Kawasaki and Fanuc. Our intention is to extend the portfolio of native robotic language functions by a number of custom-vision-related commands. This way the integrator can maintain his own working habits and easily adopt benefits of 3D vision systems in robotic applications.
We will open beta-testing and a training program for selected partners soon.
Stay tuned for more information.