AI Vision Inspection System: FA VISION

This application case features the Magenta Robotics FA VISION, a system designed for parts tending and inspection. Traditional production lines often suffer from high claim rates due to damaged, missing, or incorrectly shipped items, as well as significant delays between assembly and final taping/packaging.

FA VISION improves these inefficient systems by using AI to rapidly recognize the position and shape of objects, even measuring 3D coordinates and volume. This minimizes human error and optimizes the entire parts tending and inspection process.

Components

Robot	Vision Inspection: FA VISION AI-based high-speed object recognition. Technology for measuring 3D position, volume, and length. Ensures operational stability and increased productivity. Robot Integrated Control Framework: MAVIZ (S/W) Proprietary robot control and simulation software. Controls and simulates robot behavior by assessing the environment and context. GUI-based programming enables robot teaching and application over 4x faster than conventional methods. Multi-Item High-Speed Image Processor: MAVIX Powered by the NVIDIA JETSON TX2 board. Integrates parallel processing technology for advanced image analysis. Feature-point-based recognition module and rule-processing engine. Recognizes over 300 types of parts using deep learning. Automated quality inspection via vision recognition.

Robot

Vision Inspection: FA VISION
- AI-based high-speed object recognition.
- Technology for measuring 3D position, volume, and length.
- Ensures operational stability and increased productivity.
Robot Integrated Control Framework: MAVIZ (S/W)
- Proprietary robot control and simulation software.
- Controls and simulates robot behavior by assessing the environment and context.
- GUI-based programming enables robot teaching and application over 4x faster than conventional methods.
Multi-Item High-Speed Image Processor: MAVIX
- Powered by the NVIDIA JETSON TX2 board.
- Integrates parallel processing technology for advanced image analysis.
- Feature-point-based recognition module and rule-processing engine.
- Recognizes over 300 types of parts using deep learning.
- Automated quality inspection via vision recognition.

Workflow

STEP 1.	Parts are fed into the system.
STEP 2.	Object recognition and position verification are performed.
STEP 3.	The robot executes parts tending (handling/loading).
STEP 4.	Final inspection is conducted.
STEP 5.	Workpieces are sorted and discharged.

※ The content provided by Marosol is protected under U.S. and Canadian copyright and intellectual property laws. Unauthorized reproduction, distribution, or use of this content is strictly prohibited.

Key Features

Improved Work Productivity

Efficiency & Accuracy: Enhances overall operational efficiency and precision.
Flexible Production: Enables the construction of adaptable production lines.
Integrated Automation: Supports seamless integration with ERP systems for comprehensive Factory Automation (FA).

Optimized Motion Execution

Smart Pick-and-Place: Executes optimal movements based on sensed real-time information.
Irregular Part Handling: Capable of loading and unloading non-standard parts using suction or electromagnetic force.

Rapid Learning and Operation

High-Speed Recognition: Features rapid object detection and processing.
Intuitive Programming: Fast learning through a user-friendly GUI-based programming interface.
Remote Control: Supports real-time remote monitoring and control.

Superior Proprietary Technology

Proven Expertise: Backed by extensive on-site experience across various industries.
In-house Solution: Utilizes MAVIZ, a proprietary integrated control and simulation software.

Implementation Results

ROI Period	24 Months
Key Metrics	Rapid, accurate object recognition and 3D position/volume measurement. Easy and fast robot teaching and deployment. Highly efficient integrated parts tending and inspection system.
Client Feedback	By simultaneously performing inspection and loading/unloading tasks for products of various shapes, we have significantly reduced defect rates and greatly improved overall productivity.

※ The content provided by Marosol is protected under U.S. and Canadian copyright and intellectual property laws. Unauthorized reproduction, distribution, or use of this content is strictly prohibited.

Please log in to view

The content provided by Marosol is protected under U.S. and Canadian copyright and intellectual property laws. Unauthorized reproduction, distribution, or use of this content is strictly prohibited.

※ The content provided by Marosol is protected under U.S. and Canadian copyright and intellectual property laws. Unauthorized reproduction, distribution, or use of this content is strictly prohibited.

Components

Workflow

Key Features

Implementation Results

AI Vision Inspection System: FA VISION

Recommended Use Cases