Image Processing Apparatus Directional Division Size Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods for defect detection in inspection objects, such as those using the 'KIZUKI' Algorithm, do not provide a clear relationship between defect features and suitable parameters like resolution and division size, leading to inefficient defect extraction or increased processing load.
Innovation Solution
An image processing apparatus and method that adjust division sizes and shift amounts based on the direction of image data acquisition, allowing for effective defect detection by setting division sizes greater in one direction than the other and/or shift amounts greater in one direction than the other, and utilizing filtering processes with directionally sized filters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If division size and shift amount are not properly adjusted according to defect features, then defect detection accuracy deteriorates, but processing load and time are unnecessarily increased
Solution Approach 1:
The patent applies parameter changes by establishing specific relationships between division size, shift amount, and defect features. The setting unit adjusts these parameters based on the type and characteristics of defects to be detected, optimizing the balance between detection accuracy and processing efficiency for different inspection scenarios
Solution Approach 2:
The patent implements preliminary action by pre-establishing the correspondence between defect features and optimal parameter settings. The setting unit determines appropriate division sizes and shift amounts before the actual defect detection process begins, based on the characteristics of the inspection object and expected defect types
2Loss of time
If division size is increased to reduce processing load, then processing time decreases, but detection precision for fine defects deteriorates
Solution Approach 1:
The patent applies local quality by setting different division sizes and shift amounts in different directional dimensions (first direction vs. second direction) based on the anisotropic characteristics of defects. This allows optimized processing time in one direction while maintaining detection precision in another direction where defects may be more subtle
3Measurement precision
If multiple parameters are adjusted to match various defect features, then detection accuracy improves, but device complexity increases
Solution Approach 1:
The patent implements universality by creating a comprehensive parameter setting mechanism that handles multiple defect types and features through a unified approach. The setting unit uses a single set of rules and relationships to determine appropriate parameters for various defect characteristics, eliminating the need for multiple separate adjustment mechanisms
Data Source
AI summary
An acquisition unit acquires image data resulting from reading an image printed sequentially in a first direction. A setting unit sets a division size in the first direction and a second direction of a division area for dividing the image data, the second direction crossing the first direction, as well as a shift amount in the first direction and the second direction of the division area. A processing unit performs a predetermined process including an averaging process on the image data in accordance with the division size and the shift amount. An extraction unit extracts a unique portion from image data on which the predetermined process has been performed. The setting unit sets (i) the division size greater in the second direction than in the first direction, and/or (ii) the shift amount greater in the second direction than in the first direction.