Film Defect Detection Using Convolutional Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual defect detection in film manufacturing is inaccurate and time-consuming due to reliance on human experience, necessitating a more efficient and automated method for identifying defects in the manufacturing process.
Innovation Solution
A film defect detection system utilizing a machine learning algorithm, specifically convolutional neural networks, to analyze images of the film and detect defects, including types, numbers, sizes, and parameters such as shape, size, and color, improving accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual defect detection is used by operators, then the system is simple to operate, but the detection accuracy is low and time consumption is high
Solution Approach 1:
The patent replaces the manual mechanical inspection system with an automated optical inspection system using a camera to capture film images and a machine learning algorithm (convolutional neural network) to automatically detect and classify defects. This substitution eliminates human subjectivity and significantly improves both detection accuracy and processing speed.
Solution Approach 2:
The system enables self-service defect detection by training the machine learning model to automatically identify and classify defects without human intervention. The convolutional neural network learns from training data and autonomously performs defect detection, reducing reliance on operator experience and minimizing time consumption.
2Measurement precision
If manual defect detection is used by operators, then the equipment complexity is low, but the detection accuracy is low
Solution Approach 1:
The patent replaces simple manual inspection with an automated system comprising image acquisition hardware and machine learning software. The convolutional neural network model, trained on defect data, provides accurate defect detection and classification, significantly improving measurement precision despite the increased system complexity.
3Measurement precision
If automated machine learning defect detection is implemented, then the detection accuracy is improved, but the device complexity increases
Solution Approach 1:
The patent implements automated defect detection using a camera-based image acquisition system and a convolutional neural network algorithm. The system captures film images, processes them through the trained machine learning model, and automatically identifies defect types, positions, and characteristics, achieving high detection accuracy through automated optical and computational methods.
Solution Approach 2:
The machine learning model performs self-service by automatically learning from training data and independently executing defect detection without continuous human intervention. The system autonomously classifies defects into categories such as bubbles, tears, and contaminants, reducing the need for complex manual operation while maintaining high accuracy.
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
The system enhances defect detection accuracy by using machine learning algorithms to classify and output information on defects, reducing human error and increasing efficiency in the manufacturing process.
Implementation Method 1
The image acquisition unit acquires an image of the film by using a light transmission technology
Data Source
AI summary
The present specification provides a film defect detection system. The film defect detection system may comprise an image acquisition unit configured to acquire an image of a film in a manufacturing process of the film; a defect detection unit configured to detect defects in the film by analyzing the acquired image of the film, using a machine learning algorithm learned to detect a defect in advance, when receiving the acquired image of the film; and an information output unit configured to output information on the defects in the film detected by the defect detection unit.