Automated Optical Inspection for Satin-Finished Surfaces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing inspection methods for detecting defects on satin-finished surfaces, such as metal components, suffer from variations in accuracy due to human error and false defect detection caused by diffuse reflection of light, leading to inefficiencies in defect identification.
Innovation Solution
An inspection apparatus that uses a combination of image pickup parts and lighting sources to acquire images from multiple angles, comparing pixel values between picked-up images and reference images to detect defects with high accuracy by identifying overlapping areas based on differences and ratios, while minimizing false defect detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If visual inspection by operators is used to detect defects on satin-finished surfaces, then human judgment can identify defects, but variations in inspection accuracy and human error cause inconsistent detection results
Solution Approach 1:
The patent replaces the mechanical visual inspection system operated by humans with an automated optical inspection system using image pickup devices, light sources, and image processing units. This substitution eliminates human variability and error, providing consistent and reliable defect detection across all inspections.
Solution Approach 2:
The inspection system performs self-validation through multiple imaging directions and reference image comparisons. The apparatus automatically identifies defects by comparing captured images against reference images and analyzing discrepancies, eliminating the need for human judgment while maintaining high detection accuracy.
2Loss of information
If light is irradiated onto satin-finished surfaces for image capture, then the surface can be visualized, but diffuse reflection causes variations in tone value and increases false defect detection
Solution Approach 1:
The patent segments the inspection process into multiple imaging steps from different directions. Instead of capturing a single image that mixes surface texture and defects, the system captures multiple images from不同角度, processes them separately, and combines the results to distinguish true defects from surface texture variations.
Solution Approach 2:
The system applies different processing strategies to different regions of the image based on local characteristics. By analyzing pixel value variations in the context of local surface properties and comparing with reference images, the system can identify true defects while tolerating normal surface texture variations.
3Measurement precision
If multiple images are captured from different directions to improve defect detection, then more information is obtained, but the inspection process becomes more complex
Solution Approach 1:
The inspection system uses a universal image processing approach that handles multiple imaging directions through a single integrated processing unit. The same image pickup device, light source, and processing algorithms are used regardless of the imaging direction, simplifying the overall system design while enabling multi-directional inspection.
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 apparatus effectively suppresses false defect detection and enhances the accuracy of defect identification on satin-finished surfaces by using a multi-angle imaging and lighting approach, ensuring reliable detection of true defects.
Implementation Method 1
since the light entering the satin-finished surfaces is reflected diffusely
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
In a first defect candidate area detected on the basis of a difference between a value of each pixel in a picked-up image and a value of a corresponding pixel in a reference image and a second defect candidate area detected on the basis of a ratio between a value of each pixel in the picked-up image and a value of a corresponding pixel in the reference image, an overlapping area is detected as a defect area. It is thereby possible to suppress detection of a false defect and detect a defect with high accuracy. In a preferable defect detection part (62), a shaking comparison part (623) detects a defect candidate area on the basis of a difference in the pixel value between the picked-up image and the reference image, and a false information reducing part (629) limits pixels to be used for obtaining the above ratio to those included in the defect candidate area. It is thereby possible to detect a defect with high efficiency.