Lighting Condition Setting for Machine Vision Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In fields like product appearance checking and facial recognition, optimizing lighting conditions for machine learning-based systems is time-consuming and requires manual adjustments, limiting efficiency and accuracy due to the need for individualized settings for various workpieces or objects.
Innovation Solution
A method that uses a machine learning model to simultaneously optimize lighting and check algorithm parameters by generating estimation images under variable lighting conditions, allowing for automated setting of optimal lighting conditions for different types of objects and reducing system load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual optimization of lighting and check algorithm is performed, then checking accuracy can be improved, but design time and labor cost increase significantly
Solution Approach 1:
The system performs self-optimization by automatically adjusting lighting parameters and check algorithm parameters through machine learning without requiring manual intervention. The machine learning model learns optimal parameters independently from captured images and their corresponding label data, eliminating the need for operators to manually regulate lighting and check algorithms.
Solution Approach 2:
The patent replaces manual mechanical adjustment of lighting and algorithm parameters with an automated machine learning-based system. The machine learning model automatically determines optimal lighting parameters and check algorithm parameters by learning from training data, substituting the mechanical manual optimization process with an intelligent automated system.
2Ease of operation
If lighting is regulated to be easily observed by operator, then ease of operation improves, but checking accuracy may not be optimal
Solution Approach 1:
The system determines lighting parameters autonomously based on checking accuracy requirements rather than operator observation preferences. The machine learning model automatically selects lighting parameters that optimize detection performance by learning from labeled training data, eliminating the compromise between operator comfort and detection accuracy.
Solution Approach 2:
The patent changes the optimization criterion from operator observation ease to checking accuracy by using machine learning models trained with label data. The system adjusts lighting parameters based on their impact on detection accuracy as evaluated by the machine learning model, fundamentally changing the optimization parameter from human-centric to performance-centric.
3Measurement precision
If multiple captured images are used for optimization, then optimization accuracy improves, but system load and processing time increase
Solution Approach 1:
The patent performs preliminary action by pre-processing captured images to generate estimation images that represent lighting conditions. These estimation images are then used for training the machine learning model, separating the image processing load from the optimization calculation load and enabling efficient handling of multiple images.
Solution Approach 2:
The patent introduces estimation images as an intermediary between captured images and optimization calculations. The estimation images serve as a simplified representation that captures essential lighting information while reducing computational complexity, acting as a mediator that enables accurate optimization with reduced system load.
Data Source
AI summary
The present disclosure provides a method, device, system and computer-program product for setting a lighting condition when an object is checked and a storage medium. The method includes that: the object is lighted by light sources capable of changing lighting parameters specifying the lighting condition when the object is captured, and the object with corresponding label data is captured by an image sensor under multiple such lighting parameters to obtain multiple captured images; estimation images of the object are generated on the basis of image data sets obtained by associating the captured images and the corresponding lighting parameters; and the estimation images and the corresponding label data are applied to learning of the machine learning model, and the lighting condition is set on the basis of a comparison result between an estimation result of a machine learning model and the label data. Therefore, operations are simplified.