Mech-Vision 1.8 Release Notes
This document introduces the new features, improvements, and resolved issues of Mech-Vision 1.8.
Mech-Vision 1.8.2 Release Notes
This document introduces the new features, improvements, and resolved issues of Mech-Vision 1.8.2.
New Features
Added the “EIH System Drift Auto-Correction” Feature
After the vision system in EIH setup runs continuously for a long time or is affected by temperature changes, the picking accuracy may decrease. To solve this problem, the Error analysis tool has integrated the EIH system drift auto-correction feature. When an accuracy issue occurs with the vision system in EIH setup, this feature automatically corrects system drift after it is deployed.
New “Solution Access Control” Feature
Mech-Vision 1.8.2 introduces the solution access control feature, which enables the management of permissions for different roles to effectively reduce the data leakage risk and enhance the solution security.
After enabling the solution access control feature, all users need to log in before accessing the solution. Only after successful login can they obtain the corresponding permissions.
Control Camera Connection Permission Based on (Computer/IPC) IP Address
After Mech-Vision is upgraded to 1.8.2 and the camera firmware is upgraded to 2.3.0, multiple computers cannot connect to the same camera at the same time, avoiding interference from other users. Mech-Vision and the following clients on the same computer can still connect to the same camera at the same time:
-
Mech-Eye Viewer
-
Mech-Eye API programs
Improvement
Improved Production Interface
Starting from Mech-Vision 1.8.2, the new solution access control will be added to the production interface. The “Role-based access control” is no longer available in the configurator. After the new solution access control is configured, it will be automatically applied to the production interface.
When a solution that has been configured with the production interface of an earlier version is opened in Mech-Vision 1.8.2, the previous access control configuration in the production interface becomes invalid.
Improved Error Analysis Tool
The error analysis tool has been improved in Mech-Vision 1.8.2. A “Check DH parameters” guide was added to the Robot page. The robot accuracy error will be introduced when the pose data transferred between the robot and the vision system is in the form of joint positions, and the DH parameters of the real robot do not match those of the simulated robot. In such a case, you can refer to this page to check the robot DH parameters.
Improved Step Library
Mech-Vision 1.8.2 improved the Step Library by removing the button on the right of the search bar that showed frequently-used Steps. It also added the “Frequently-Used Step” and “Frequently-Used Procedure” categories at the top of the Step Library, which makes it easier to find the Steps you need.
Improved Steps
Mech-Vision 1.8.2 improved some Steps. See the following for details.
| No. | Step | Improvement | Description |
|---|---|---|---|
1 |
Added the “Auto-Calculate Expected Point Count” parameter. |
This parameter is used to determine whether to automatically calculate the number of points in the down-sampled point cloud. Once this option is selected, the number of points in the down-sampled point cloud will be calculated automatically. If it is not enabled, you need to manually specify the “Expected Point Count of Sampled Model” parameter. |
|
2 |
Added the “Result Validation Degree” parameter. |
This parameter is used to select the degree of strictness applied when validating the matching results. |
|
3 |
External triggering is supported. |
The data acquisition can be triggered by an external signal. |
|
4 |
Improved some parameters. |
|
|
Improve graphics card requirements. |
When using this Step, it is recommended to use a graphics card between NVIDIA GTX 1660 Ti and RTX 3060 Ti (including GTX 1660 Ti and RTX 3060 Ti). Graphics cards below NVIDIA GTX 1650 Ti are no longer supported. |
||
5 |
Added the “Inference Mode” parameter. |
This parameter is used to select the deep learning inference mode, including GPU and CPU. |
|
6 |
The picking applications in supermarket, logistics, and cables scenarios are no longer supported. |
Starting from Mech-Vision 1.8.2, this Step no longer supports picking applications in supermarket, logistics, and cables scenarios. |
|
Added the dynamic port function. |
From Mech-Vision 1.8.2, the “Predict Pick Points V2” Step displays the input and output ports dynamically. This Step displays different input and output ports according to the picking configuration folders in different scenarios. |
Improved Extrinsic Parameter Calibration Report
The extrinsic parameter calibration report has been improved in Mech-Vision 1.8.2. The information in the extrinsic parameter calibration report was simplified, highlighting important information and improving the readability of the extrinsic parameter calibration report.
Resolved Issues
The following issues have been resolved in Mech-Vision 1.8.2:
-
When the zoom percentage of the laptop was 125%, the “Translation Adjustment” parameter panel in the pose adjustment tool was too small to facilitate parameter adjustment.
-
In some cases, the solution could not be opened normally after the production interface was configured.
-
The “Capture Images from Camera” Step could not automatically reconnect to the camera to capture images after the camera was disconnected.
-
After the “Capture Images from Camera” Step was used to run the project in virtual mode and the project was saved, the “Camera Calibration Parameters” parameter was empty when a real camera was used.
-
When the number of cameras in a solution was equal to or exceeded 5, the connection statuses of only 5 cameras were displayed in the status area in the production interface, and the connection statuses of other cameras were displayed as empty.
-
After the camera was disconnected, the camera status in the production interface status area was still displayed as “Connected”.
-
The timeout for obtaining Mech-Viz data displayed in the window was 10 seconds, but it was actually 5 seconds.
-
In the pre-calibration configuration for gantry robots, after selecting “No movement affecting the camera position,” the robot flange pose still needed to be input when running the project.