A computer vision-based system is made of many different components to be selected case by case during the feasibility process (camera, optics, illumination, computer, controllers, software, etc.)
TCMM360-Ai1 is a total vision solution for inspection and quality control of objects from multiple views using a single shot.
TCMM-360-26M-02 Using 30° Mirror Configurations
TCMM-360-26M-02 Using 45° Mirror Configurations
Telecentric optics allow inspection and measurement. Featuring ROI cutting and geometrical transformations of the sample in the integrated software.
| Rule-based Preprocessing Tools | These include image enhancement, geometric transformations of the ROI, Color Space Conversion, Morphological operations, and others |
|---|---|
| System Architecture Builder | A visual no-code workflow builder for selecting input channels (such as cameras, on-device offline folders, or external servers), applying preprocessing, cutting ROIs, and assigning classes and models. It supports multi-model inference on a single picture or individual channels of an RGB camera, for example. |
| Communication Management | Manage through hardware GPIO and Serials to directly control sensors and actuators or use Ethernet for communication with external devices like PLCs, computers, or other SmartPC-Ai1 units. |
| Illumination Control | The software includes an integrated high-power 48V, 4-channel controller with independent settings for pulse duration and an auto-strobe mode that emulates continuous acquisition with all the advantages of the pulsed mode. |
| Management Tools | Each software version comes with remote desktop, model backup, firmware updates via FTP or USB, and functions for saving images to external drivers or NAS. Subscription to the System Integration or Far Island Partner program unlocks additional tools for online remote management (over the internet or intranet), mass deployment of models to multiple machines, and more. |
| Deep Learning Tools | "The highlight is the AI toolset, which is GUI-based and streamlines the workflow in 4 steps: Data acquisition, Labeling, Training, and Model deployment/ prediction.Training can be performed on Far Island's cloud servers or directly on the edge device.Training directly on the edge device is slower but absolutely unique in the whole computer vision industry.Subscriptions offer additional options such as purchasing training time, popular choice for end -users, or dedicated GPU or Servers for SI and partners' Companies who wish to be fully independent or generate revenue by offering training subscriptions to their clients." |
The standard software included in the SmartPC-Ai1 has on-device everything needed to develop the vision solution on any possible optical system. CCTV, Telecentric, Multi-Mirrors, Hypercentric, Catadioptric, pinhole, probes, are selectable from the GUI. This includes adjustments for the ROI size, position, the number of mirrors, unwrapping tools, and more. Additionally, the GUI integrates all necessary hardware peripherals and communication interfaces for interacting with the external world, ensuring a ready-to-use experience.
The fully GUI-based approach requires no coding skills and empowers anyone with state-of-the-art tools to develop a Deep Learning-based inspection solution in a few hours.
SmartPC-Ai1 Standalone
SmartPC-Ai1 Multi-Mirror (TCMM)
SmartPC-Ai1 Catadioptric (CDAA)
SmartPC-Ai1 Custom Solutions
SmartPC Bench-Ai1 (Metrology)
