Visual Feedback for Inspection Algorithm Filter Settings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inspection systems face difficulties in setting up and evaluating difference filters and inspection algorithms due to the inability to directly observe their effects on processed test and reference images, leading to avoidance by application engineers and customers.
Innovation Solution
A system providing visual feedback through a user interface that processes sample images using a selected inspection algorithm and/or difference filter, offering visual representations such as signed difference images, side-by-side views of original and filtered images, and binary scatter plots to facilitate the selection of optimal settings for improved defect detection and separation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If difference filters and inspection algorithms are used to process sample images, then defect detection and separation are improved, but the difficulty of setup and evaluation increases
Solution Approach 1:
The system applies feedback by displaying visual representations (difference images, scatter plots, histograms) that show the effects of filter and algorithm settings in real-time. This allows users to observe how different parameters affect defect detection and adjust settings accordingly, transforming an otherwise opaque process into one with immediate visual feedback loops.
Solution Approach 2:
Visual representations serve as intermediaries between the complex filter/algorithm processing and the user. Instead of directly exposing users to complex parameter settings and mathematical transformations, the system mediates through visual outputs like difference images and scatter plots that bridge the gap between technical processing and user comprehension.
2Measurement precision
If complex inspection algorithms are applied to sample images, then defect separation capability is improved, but the ability to directly observe processing effects is lost
Solution Approach 1:
The system segments the processing workflow into distinct visualizable stages: raw image input, difference image generation, filter application, scatter plot generation, and histogram display. Each stage produces a visual representation that preserves information about the processing effects, allowing users to trace how defects are separated at each step rather than losing visibility in a black-box process.
Solution Approach 2:
The system utilizes color and intensity variations in visual representations to encode information about processing effects. Different colors and intensity levels in difference images, scatter plots, and histograms represent different defect characteristics and processing outcomes, making abstract algorithm effects visible through visual color and brightness changes.
Data Source
AI summary
The disclosure is directed to providing visual feedback for inspection algorithms and difference filters used to process test and reference images from an inspection system. A user interface may be configured for displaying information and accepting user commands. A computing system communicatively coupled to the user interface may be configured to receive at least one set of test and reference images collected by the inspection system. The computing system may be further configured to provide at least one visual representation of the test and reference images via the user interface to show effects of an inspection algorithm and/or difference filter.