Introduction

This module is used to segment the contour of target objects and output the corresponding labels of the classes.

Applicable Scenarios

Depalletizing: Applicable to scenarios where cartons, boxes, or sacks need to be depalletized from the pallets and then placed elsewhere (such as a bag break station or conveyor belt).

../../../_images/sack.png
../../../_images/instance_segmentation_31.png

Machine tending: Applicable to machine tending of complex workpieces, structural parts, irregular parts in automobile, steel, machinery, and other industries.

../../../_images/rotor.png
../../../_images/instance_segmentation_21.png

Product picking: Applicable to batch picking, discrete order picking, etc., in the e-commerce industry. This module is capable of processing various product packaging such as airbags, transparent packaging, aluminum cans, packaging of irregular shapes, etc.

../../../_images/goods.png
../../../_images/instance_segmentation_introduce.png

Express parcels: Suitable for detecting parcels of different shapes such as soft packages, postal envelopes, express cartons, padded envelopes, etc.

../../../_images/package.png

General Workflow

../../../_images/application_flow3.png

Tips

The performance of the trained model depends on the followings.

  1. The quality of the captured images.

  2. The quality of the datasets.

  3. The quality of labeling.