Shape Recognition and Sorting Automation using Neuromeka Indy7 and IndyEye

Industry > Metal·PlasticsIndustry > ElectronicsApplication > ClassificationSector > Manufacturing

This process is a demo for shape recognition and product sorting automation utilizing 2D vision and a deep learning algorithm (Instance Segmentation).

Because it performs tasks by recognizing product positions and types in real-time using 2D vision, it can automatically recognize and work even if product positions or types change.

Components

Robot	Robot Indy7: 6-axis collaborative robot, payload 7kg, maximum reach 1.3m, weight 28kg, repeatability: ± 0.1 mm

Workflow

STEP 1.	Workpiece moves to a designated position on a conveyor belt.
STEP 2.	Shape recognition of the object through deep learning.
STEP 3.	Pick and place of the workpiece

※ 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

Easy and simple robot teaching

: Reduced teaching time through vision calibration and teaching automation utilizing deep learning

Excellent recognition for various objects by recognizing objects based on deep learning

Automatic learning of optimal pick positions according to product shapes

Compact and simple installation

: No need for separate feeder production as there is no need to fix inputs in a precise position

Possible to work by simply installing the robot in the space where the worker used to work

Implementation Results

ROI Period	Null
Key Metrics
Client Feedback	Thanks to IndyEye, which uses deep learning and vision sensors, fast and simple teaching is possible, and picking and placing can be performed by recognizing the shape of products in real-time.