Adaptive Object Sorting System with Dynamic Lighting and Imaging Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional product inspection and sorting systems are specialized and cannot effectively sort diverse types of objects, as they lack adaptable inspection and sorting rules, making them unsuitable for various products such as timber, rice, and fruits when used with the same settings.
Innovation Solution
A system comprising a light source, image capturing device, and a controlling and processing device with self-learning decision units that adjust parameters and classifiers based on object images, enabling adaptive sorting of objects into normal and defective groups, applicable to any type of object.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional product inspection and sorting systems use fixed inspection and sorting rules, then they can effectively sort specific object types (e.g., timber, rice, or fruits), but they cannot adapt to sort diverse types of objects
Solution Approach 1:
The system dynamically adjusts inspection and sorting rules based on the type of object being processed. The control device receives object type information and automatically selects or generates appropriate inspection rules and sorting criteria, transforming the fixed system into a dynamic one that adapts to different objects without requiring manual reconfiguration.
Solution Approach 2:
The system changes inspection parameters and sorting rules according to the object type. Different objects (timber, rice, fruits, etc.) have different optimal inspection parameters and sorting criteria, which the system adjusts automatically based on the detected object type, enabling versatile operation while maintaining optimal performance for each object category.
2Measurement precision
If the system uses standardized inspection rules for all objects, then the system structure remains simple, but the sorting accuracy and quality control deteriorate for different object types
Solution Approach 1:
The inspection rules are made dynamic rather than static. The control device automatically selects or generates appropriate inspection rules based on the object type, allowing the system to achieve high sorting accuracy for each specific object category while maintaining a unified system architecture that doesn't require separate dedicated systems for each object type.
3Productivity
If manual inspection rules are used for each object type, then sorting accuracy is maintained, but the automation level and productivity decrease
Solution Approach 1:
The system performs self-configuration by automatically selecting or generating appropriate inspection rules and sorting criteria based on the detected object type. This self-service capability eliminates the need for manual rule configuration for each object type while maintaining high sorting accuracy, thereby achieving both automation and precision.
Solution Approach 2:
The system uses feedback from object type identification to automatically adjust inspection rules and sorting parameters. The control device receives information about the object type and uses this feedback to select or generate appropriate inspection rules, creating a closed-loop system that maintains high accuracy while fully automated.
Data Source
AI summary
A system for sorting moving objects is disclosed. The system comprises a light source, an image capturing device, a controlling and processing device, and an object sorting device. Particularly, the controlling and processing device is configured to decide a first setting parameter so as to apply a parameter adjustment to the light source, and is also configured to decide a second setting parameter so as to apply an parameter adjustment to the image capturing device. After deciding an object classifier based on the first setting parameter, the second setting parameter, and object images received from the image capturing device, the object sorting device is controlled to apply an object sorting process to the of objects that are delivered by the belt conveyor, thereby sorting the objects into at least two object group consisting of a normal object group and a defective object group.