Optical Inspection and Sample Sorting for Target Quality Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automated quality inspection systems lack efficiency and accuracy in detecting defects in products, particularly for items like tree nuts and tablets, and require significant user training and maintenance.
Innovation Solution
A computing device with a processing circuit, memory unit, and communication port, which receives target and measured quality values to generate a source control instruction, directing a source selecting device to sort samples based on their quality characteristics, using optical inspectors and neural networks for defect detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated quality inspection systems are implemented, then productivity and defect detection capability are improved, but device complexity and initial cost increase
Solution Approach 1:
The patent replaces complex mechanical inspection systems with optical inspection devices that use light-based detection. The optical inspector captures images of products, and neural networks process these images to detect defects, eliminating the need for complex mechanical measurement and inspection mechanisms while maintaining high detection capability.
Solution Approach 2:
The system creates digital copies (images) of physical products using optical inspectors. These digital representations are then analyzed by neural networks, allowing complex inspection tasks to be performed on simplified digital data rather than requiring complex physical inspection mechanisms.
2Measurement precision
If traditional quality inspection methods are used, then device complexity is reduced, but measurement precision and detection accuracy deteriorate
Solution Approach 1:
Traditional mechanical measurement devices are replaced with optical inspection systems combined with neural network analysis. The optical inspector captures high-resolution images, and the neural network provides intelligent defect detection, achieving superior measurement precision without requiring complex mechanical measurement mechanisms.
Solution Approach 2:
The patent introduces an intermediary layer (neural network) between the optical inspection device and the defect detection process. This intermediary processes the visual data intelligently, providing high detection accuracy while keeping the physical inspection device relatively simple in structure.
3Productivity
If automated sorting systems are implemented, then productivity is improved, but ease of operation and user training requirements worsen
Solution Approach 1:
The sorting system operates autonomously based on real-time inspection data. The neural network automatically analyzes product images, determines quality categories, and controls the sorting mechanism without requiring manual intervention or specialized operator training. The system serves itself by making all decisions based on the inspection data it collects.
Solution Approach 2:
The system implements a closed-loop feedback mechanism where inspection results directly control sorting actions. The neural network continuously analyzes inspection data and adjusts sorting decisions in real-time, creating an autonomous system that requires minimal human input or training while maintaining high sorting efficiency.
Data Source
AI summary
A process includes receiving a target quality value, receiving a measured quality value, receiving a source quality value, and sending a source control instruction. The source control instruction is based at least in part on the target quality value, the measured quality value, and the source quality value. The target quality value, the measured quality value, the source quality value, and the source control instruction are communicated via the communication port. The measured quality value is generated by an inspection device configured to inspect a sample. The source quality value is associated with a quality level of a first group of samples. The target quality value indicates a desired quality value of an output group of samples. The source control instruction causes a source selecting device to select one of a plurality of groups of samples, each group having identified quality characteristics.