Workpiece Sorting System Using Image and Weight Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional workpiece sorting systems face inefficiencies in sorting various types of workpieces with different weights, particularly when they are conveyed randomly, as they require multiple sorting destinations and can be space-intensive, and may incorrectly classify workpieces based on image analysis alone.
Innovation Solution
A workpiece sorting system that uses a conveyor, cameras for image capture, robots with weight measurement capabilities, and a control system to classify workpieces into large groups based on images and allocate them to robots, which then sort into smaller groups based on weight measurements, allowing for efficient sorting and correction of misclassifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If workpieces are sorted using only image analysis by a single robot, then the system is simple, but classification accuracy deteriorates due to incorrect classification of workpieces with similar appearances but different weights
Solution Approach 1:
The sorting process is segmented into two stages: first, a single robot performs preliminary classification into large groups based on image analysis; second, multiple robots perform detailed sorting into small groups based on weight measurement. This segmentation allows each robot to specialize in specific tasks, improving overall accuracy while managing system complexity.
Solution Approach 2:
Weight measurement acts as an intermediary verification step between image-based preliminary classification and final sorting. The weight measurement part measures the weight of workpieces in large groups, and the control part uses this information to accurately classify workpieces into small groups, correcting potential misclassifications from image analysis alone.
2Adaptability or versatility
If multiple sorting destinations are provided for various types of workpieces, then sorting completeness is improved, but the system becomes space-intensive
Solution Approach 1:
The sorting destinations are segmented into two hierarchical levels: large groups (e.g., light, medium, heavy) and small groups (sub-categories within each weight class). This segmentation allows workpieces to be first directed to appropriate large group areas, then further sorted into small groups, reducing the need for numerous separate sorting destinations while maintaining comprehensive sorting capability.
Solution Approach 2:
The sorting system adds a temporal dimension to the sorting process by using sequential processing: first spatial classification into large groups, then weight-based classification into small groups. This two-dimensional sorting approach (spatial + weight) allows efficient use of sorting destination space while handling diverse workpiece types.
3Productivity
If a single robot handles all sorting tasks, then the system is simple, but processing speed deteriorates when handling large varieties of workpieces
Solution Approach 1:
The workload is segmented among multiple robots based on their specialized functions: one robot handles image-based preliminary classification, while other robots handle weight-based detailed sorting. This segmentation allows parallel processing of different task types, increasing overall processing speed without requiring every robot to perform all sorting operations.
Solution Approach 2:
The control part serves as a universal coordinating system that manages multiple robots, allocating workpieces to appropriate robots based on their capabilities and current workload. This multi-functional control approach allows the system to efficiently handle diverse workpiece varieties while maintaining manageable robot complexity.
Data Source
AI summary
Provided is a sort workpiece sorting system including: a conveyor; a conveying position detection part for workpieces; a camera configured to photograph the workpieces; robots provided with a hand disposed along a conveying direction and capable of ejecting workpieces on the conveyor, and a weight measurement part; a control part configured to control each robot; and a large group classifying part configured to classify the workpieces into large groups allocated to the robots based on an image of the workpieces. The control part includes a visual tracking part configured to cause the robots to handle the workpieces while following an operation of the conveyor based on the image of the workpieces and a conveying position. The visual tracking part causes the visual tracking part to eject the classified workpieces are ejected from the conveyor by the hand, and the workpieces are sorted for each group based on the measured weight.