Windshield Optical Defect Detection Using Ghost Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting optical defects in glass windshields, particularly threads, are ineffective, leading to undetected defects that can cause customer rejection and production losses, as they are often overlooked by current inspection systems.
Innovation Solution
A computer-implemented method using angular distance maps and image processing to detect optical defects, specifically threads, by analyzing the angular distances and shape parameters of periodical patterns through the windshield, distinguishing between genuine defects and other elements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional deflectometry-based inspection systems are used to measure optical distortion, then the measurement process is efficient and covers the whole windshield, but peculiar optical defects called threads are not detected because no trace of these defects remains in the optical distortion map
Solution Approach 1:
The patent segments the inspection approach into two distinct parts: (1) conventional deflectometry for measuring overall optical distortion, and (2) a supplementary method analyzing specific image features (dot positions, shapes, intensities) to detect threads. This segmentation allows each method to focus on what it does best while combining them provides comprehensive defect detection capability.
Solution Approach 2:
The patent introduces an intermediary analysis layer that processes the raw image data from the deflectometry system. Instead of directly using only the optical distortion map, the system examines intermediate image features (position, shape, intensity of transmitted dots) that serve as indicators of thread defects. This intermediary analysis reveals defects that would otherwise be invisible in the final distortion map.
2Device complexity
If inspection systems rely solely on optical distortion maps from deflectometry, then the system complexity is low and implementation is simple, but the ability to distinguish genuine optical defects from masking elements (gasket, enamel, ink marks) is lost
Solution Approach 1:
The patent applies local quality analysis by examining specific local features of transmitted image elements (dot position, shape, intensity) rather than relying on global optical distortion characteristics. Different defect types produce distinct local feature patterns, allowing the system to differentiate between genuine optical defects and masking elements based on their unique local signatures.
Solution Approach 2:
The patent changes the parameters being analyzed from global optical distortion values to local image features including dot position coordinates, shape parameters, and intensity values. This parameter transformation enables more reliable defect identification because these local parameters exhibit characteristic patterns for different defect types, improving the system's ability to distinguish genuine defects from masking elements.
3Productivity
If current inspection methods are used, then production flow is maintained with simple inspection, but undetected thread defects lead to customer rejections and production losses
Solution Approach 1:
The patent implements preliminary action by detecting and flagging thread defects during the manufacturing inspection phase using the enhanced image analysis method. By identifying defects before the windshield leaves the production line, the system prevents defective products from reaching customers, thereby maintaining production flow while ensuring quality assurance through early defect detection.
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
Effectively detects optical defects like threads that are undetected by conventional systems, ensuring higher quality control and reducing production losses.
Implementation Method 1
a map of angular distances between a primary and a ghost image of a periodical pattern viewed through a glass windshield and an image of the periodical pattern through the glass windshield are acquired with an inspection system
Implementation Method 2
a map of angular distances between a primary and a ghost image of a periodical pattern viewed through a glass windshield
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A computer implemented method for detecting optical defects within a glass windshield. The method takes as input a map of angular distances between primary and ghost images of a periodical pattern viewed through said glass windshield, an image of the periodical pattern viewed through said glass windshield, and at least one shape parameter of the individual features of the periodical pattern. The method provides as output an image of optical defects within the glass windshield.