Mech-MSR Release Notes

Supports capturing images with third-party industrial 2D cameras and controlling lighting with the Light Source Step. See the 2D Camera Step for details.

Supports distortion calibration for 2D cameras to correct geometric distortions caused during lens imaging. See Distortion Calibration (2D Camera) for details.

Adds a large number of Steps for 2D scenarios, covering image preprocessing, locating, recognition, measurement, and post-processing, which can be used to build complete 2D inspection and measurement workflows. See the relevant Steps under the 2D category in the Step Library for details.

Added multiple algorithm Steps

Supports the feature region group

Deep learning: Deep Learning Result Parser

Judgment & control: Logical Judgment, Trigger Control by Flag

File reading and saving: Read Poses from File, Save Data to Files

Data processing: Numeric Operation, Convert Data Type

Quick setup: Easy Create Number List, Easy Create Poses, Easy Create String List

Data communication: Receive Data, Parse Data, Assemble Data, Trigger Project, Send Data

Supports configuring multiple custom communication services, including TCP Server, TCP Client, Siemens S7 Client, Modbus TCP Slave, and Mitsubishi MC Client.

Supports unifying communication parsing, project triggering, measurement execution, and result return into one configurable workflow, without developing separate scripts or Adapter for each project.

Supports building communication and measurement projects. Data processing and result return can be completed through Steps such as Receive Data, Parse Data, Logical Judgment, Trigger Project, Assemble Data, and Send Data.

Supports managing custom communication services and configuring communication trigger rules: For TCP Server and TCP Client services, receiving data can trigger execution. For PLC services, the corresponding project can be triggered based on register variable values. See Communication Configuration and Custom Communication Mode for details.

Tilt calibration: Supports correcting sensor head mounting tilt automatically or manually to reduce data distortion and measurement errors caused by tilt. See Tilt Calibration (Auto) and Tilt Calibration (Manual) for details.

Stitching calibration: Supports segmented data acquisition on a calibration target to achieve accurate stitching of multiple data segments. Suitable for calibration scenarios involving large or elongated objects. See Stitching Calibration for details.

The judgment result encoding is adjusted to OK = 1 and NG = 0.

The normal status code is adjusted from 0 to 1, making it easier to distinguish a successful status from other default values in practical applications.

The Deep Learning Model Package Inference Step now supports 2D image-based deep learning scenarios, including image classification, object detection, defect segmentation, instance segmentation, fast positioning, text detection, text recognition, and unsupervised segmentation. See Deep Learning Model Package Inference for details.

In 3D scenarios, the Deep Learning Model Package Inference Step can now use single model packages for fast positioning and multi-model packages for inference to meet more complex inspection workflow requirements.

Inference configuration has been optimized. The Inference Configuration Tool can now be used to flexibly adjust the post-processing parameters of the Deep Learning Model Package Inference Step, validate multi-model package configurations, and support parameter group management and visualized parameter tuning.

The deep learning model package management tool has been upgraded to support unified management of single model packages and multi-model packages. You can view model package types, projects that use them, and model package status, and you can also release or delete model packages for more flexible resource management. See Deep Learning Model Package Management Tool for details.

The Data Viewer has been upgraded to support switching between 2D Data View and 3D Data View when viewing visualized data. See Data Viewer for details.

A data details icon is added to the Output Results panel, allowing you to quickly view more result information such as data type, data count, and data dimensions. See Result Display Panel for details.

The Global Variable Management tool has been upgraded. In addition to viewing saved global variables, it now supports centralized creation and management of global variables in the current solution, including adding, deleting, clearing, drag-and-drop sorting, and resetting variable IDs.

Supports unified configuration of data type, initial value, current value, bound parameters, and communication switch for global variables, and supports assigning global variable values to Step parameters.

The workflow of the Global Variable Step has been improved. You can open the global variable management tool directly from the Step. For queue-mode global variables, the Retrieve All parameter can also be used to retrieve all data at once.

Added the View Parameters feature. You can view and adjust measurement item parameters in the current project in the production interface, and parameter import and export are also supported.

The display window configuration capability in the production interface has been upgraded. Background images now support not only surface data, but also profiles and 2D images. See Configure Display Windows for details.

The overlay capabilities of display windows are further enhanced. On the background image of a 2D image and surface data, poses can now be overlaid. In addition, the maximum number of overlay items such as features, images, and text is increased, allowing inspection and measurement results to be displayed more flexibly.

The 3D Laser Profiler Step has been upgraded so that results of tilt calibration, S/Z scan stitching calibration, and N-point calibration can be applied directly during acquisition, enabling more flexible adaptation to scenarios such as segmented scanning and machine coordinate mapping.

The Extract Surface Section Step has been upgraded with a new Horizontal centerline sectioning mode, which allows a profile to be extracted directly based on the horizontal centerline of the surface data bounding box.

The Fit Circle to Surface Edge Step has been upgraded with an outlier handling option. When least squares circle fitting is used, outliers can be handled automatically, or the outlier fraction can be set manually, improving usability in complex edge scenarios.

The height measurement logic of the Measure Surface Track Step has been optimized. Height can now be calculated based on an average point, fitted line, reference plane, or with no reference, making it more suitable for complex scenarios such as tracks on sloped surfaces.

The Measure Cylinder Step has been upgraded to simplify fitting parameter configuration, and it now provides cylinder geometry output for direct use of fitting results in subsequent workflows.

The Detect and Measure Hole on Surface Step has been upgraded to support more flexible reference plane configuration: In addition to fitting a reference plane from the reference region within the Step, you can also directly input an external reference plane. It also adds circle geometry output for subsequent processing and display.

The Extract Alignment Parameter Group Step has been upgraded to support both 2D and 3D alignment modes, allowing alignment parameter groups for different scenarios to be extracted as needed.

The Save Surface Data Step in version 2.1.1 has been upgraded to the Save 3D Data Step. In addition to surface data, it also supports saving point clouds and profiles, and lets you choose the appropriate format for different data types.

The Data Storage feature has been upgraded with new options to create subfolders by parameter recipe or by OK/NG result, making it easier to save data by production condition.

The Statistical Report feature has been upgraded. In addition to querying by date range, it also supports querying by the most recent number of project runs, making it easier to quickly view statistics from recent production results.