Blister Pack Inspection Using Neural Network Shading Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing blister pack inspection methods fail to accurately detect formation defects in the side portions of pocket portions, particularly when the bottom portion has uneven wall thickness distribution or a complicated shape, leading to potential gas barrier property reduction.
Innovation Solution
An inspection device using a neural network-based system that irradiates the container film with a predetermined electromagnetic wave, extracts shading pattern data from the bottom portion, and reconstructs it to compare with actual data for accurate quality judgment of the side portion, enabling detection of subtle defects and improving productivity by imaging only the bottom portion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the bottom portion has uneven wall thickness distribution or a complicated shape, then the bottom portion wall thickness measurement may appear normal, but the side portion wall thickness becomes inaccurate leading to formation defects
Solution Approach 1:
The patent creates a reference image of a normal pocket portion and uses image processing to generate a virtual reference image that accounts for the specific geometry and wall thickness distribution of the bottom portion. This virtual reference image serves as a customized template for comparing against actual side portion images, enabling accurate defect detection even when the bottom portion has uneven wall thickness or complicated shapes.
Solution Approach 2:
The patent transforms the inspection approach by changing from direct side portion measurement to bottom portion measurement with virtual reference generation. By adjusting the reference image parameters based on the specific bottom portion characteristics (shape, wall thickness distribution), the system adapts the inspection criteria to match the actual geometry, thereby improving measurement accuracy for side portions with varying wall thicknesses.
2Manufacturing precision
If inspection is performed on the side portion directly, then formation defects can be detected, but the inspection time and complexity increase significantly
Solution Approach 1:
The patent extracts the inspection target from the side portion to the bottom portion, which has more favorable imaging characteristics. By measuring the bottom portion wall thickness and using it to generate a virtual reference image, the system avoids the time-consuming direct side portion inspection while maintaining defect detection capability through the correlation between bottom and side portion wall thicknesses.
Solution Approach 2:
The patent performs preliminary image processing to generate a virtual reference image based on the bottom portion characteristics before conducting the actual defect detection. This pre-processing step creates an optimized reference that accounts for the specific geometry, enabling faster and more accurate comparison with actual side portion images without requiring complex real-time adjustments during inspection.
3Productivity
If conventional image processing is used for side portion inspection, then defects can be detected, but the inspection accuracy is insufficient for complicated bottom portion shapes
Solution Approach 1:
The patent applies local quality by generating a customized virtual reference image that specifically matches the local characteristics of each bottom portion (shape, wall thickness distribution). Instead of using a generic reference image, the system creates a locally optimized reference that accounts for the specific geometry of each pocket portion, thereby improving measurement precision for side portions with complicated shapes while maintaining inspection speed through automated image processing.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This solution enhances the accuracy of detecting formation defects in the side portions and increases the speed of inspection, reducing the time and labor required for each pocket portion, while preventing defective blister packs from being filled and improving overall productivity.
Implementation Method 1
an irradiation unit (i.e., an illumination device) configured to irradiate a container film with the pocket portion formed therein with a predetermined electromagnetic wave
Data Source
AI summary
An inspection device inspects a formation state of a pocket portion formed in a container film of a blister pack and includes: an illumination device that irradiates a container film including the pocket portion with a predetermined electromagnetic wave; an imaging device that takes an image of at least the electromagnetic wave transmitted through a bottom portion of the pocket portion and obtains image data; a control device that extracts, based on the image data, shading pattern data corresponding to a shading pattern occurring in the bottom portion of the pocket portion by irradiation with the electromagnetic wave; a storage that stores a neural network and a model, the model being generated by learning of the neural network using, as learning data, only shading pattern data of a pocket portion without any formation defect among the extracted shading pattern data.