lntevega_Galatek

We use cookies to improve your experience on our websites and to make visit statistics.

InteVega

With artificial intelligence and vision applications at its core, Galatek innovatively developed the InteVega AI vision platform. The platform deeply integrates intelligent control, machine vision and deep learning technologies, integrates data flow, 2D and 3D image processing, multi-modal large model, AIGC sample generation, small sample learning and other technologies, and efficiently builds flexible and customizable AI vision software for a variety of application scenarios in a zero-threshold, one-stop way.

InteVega Zero-threshold One-stop AI Vision Platform

●Based on data flow engine and more than 500 self-developed 2D/3D vision +AI operators, the platform covers measurement, guidance, detection, recognition and other machine vision application scenarios;

●Based on the self-developed data flow engine and flow chart editor, the algorithm visualization, combination, drag and drop development is realized.

●Support image acquisition, communication between devices, motion control, production interface customization and other key vision development functions;

●It covers 10 types of algorithm scenarios such as detection, positioning, semantic segmentation, instance segmentation, classification, OCR/OCV, and unsupervised anomaly detection, and breaks through the whole process of data acquisition, labeling, grouping, training, model pruning quantization acceleration, and embedded vision process inference, realizing the zero code development and going on-line of AI algorithm model.

Core module

Vision Software Platform

Providing camera management, light source management, PLC management, user rights management, production layout settings, production statistics and reports, logs, system settings and other basic vision modules. Supports 90 days *24 hours running cycle.
Vision Algorithm Data Flow

Provide 500 self-developed operators, including 2D+3D+A+ script + program control + logic judgment. Build vision algorithms based on flow chart editor, enabling zero-code development and debugging of vision algorithms..
AI Algorithm Platform

For 10 types, more than 50 kinds of SOTA-level AI algorithms, targeted optimization for machine vision scenes are performed. Providing a whole process from image preprocessing, data annotation, model tuning, model training, inference testing to online deployment, 0 threshold AI algorithm development platform.

Vision Software Platform

Providing camera management, light source management, PLC management, user rights management, production layout settings, production statistics and reports, logs, system settings and other basic vision modules. Supports 90 days *24 hours running cycle.

Vision Algorithm Data Flow

Provide 500 self-developed operators, including 2D+3D+A+ script + program control + logic judgment. Build vision algorithms based on flow chart editor, enabling zero-code development and debugging of vision algorithms..

AI Algorithm Platform

For 10 types, more than 50 kinds of SOTA-level AI algorithms, targeted optimization for machine vision scenes are performed. Providing a whole process from image preprocessing, data annotation, model tuning, model training, inference testing to online deployment, 0 threshold AI algorithm development platform.

Features

Multi-modal Large Model

Galatek AI is the first to introduce self-developed semantic image large model technology to the field of machine vision object detection and recognition. Combined with cue word engineering, zero-sample/low-sample learning is realized in a way that is closer to human sensory description.
AIGC

Galatek AI is the first to introduce self-developed AIGC technology to the field of machine vision defect detection. Through intuitive commands, functions such as defect generation, defect background change, defect shape change, etc. can be achieved by specifying category and position, which greatly alleviates the problem of difficult NG sample collection and improves the precision of industrial quality inspection.
One Click Automatic Calibration

Support N point calibration, rotating center calibration and multi-camera correlation calibration. Provide standard command set. After the communication configuration is completed, one-click automatic calibration and zero threshold development can be realized, which effectively improve calibration efficiency.
Galatek Flow

Based on the data flow engine technology to realize the modular encapsulation of algorithms, the wired-based operation can complete the application configuration with zero threshold, and achieve the seamless integration of 2D+3D+AI. Support parallel operation and logical operation to improve the execution efficiency. Support intermediate process recording, analysis and playback.

Multi-modal Large Model

Galatek AI is the first to introduce self-developed semantic image large model technology to the field of machine vision object detection and recognition. Combined with cue word engineering, zero-sample/low-sample learning is realized in a way that is closer to human sensory description.

AIGC

Galatek AI is the first to introduce self-developed AIGC technology to the field of machine vision defect detection. Through intuitive commands, functions such as defect generation, defect background change, defect shape change, etc. can be achieved by specifying category and position, which greatly alleviates the problem of difficult NG sample collection and improves the precision of industrial quality inspection.

One Click Automatic Calibration

Support N point calibration, rotating center calibration and multi-camera correlation calibration. Provide standard command set. After the communication configuration is completed, one-click automatic calibration and zero threshold development can be realized, which effectively improve calibration efficiency.

Galatek Flow

Based on the data flow engine technology to realize the modular encapsulation of algorithms, the wired-based operation can complete the application configuration with zero threshold, and achieve the seamless integration of 2D+3D+AI. Support parallel operation and logical operation to improve the execution efficiency. Support intermediate process recording, analysis and playback.

Application Scenarios

Detection

Strong detection ability for defects with complex background and low contrast, covering scratches, cracks, hidden cracks, foreign matters, dirt, missing and other scenes of industrial defect detection needs. Effectively addresses small sample detection problems, suitable for multi-batch, small-batch flexible production scenarios.