Defect Minimap Comparison for Rotation-Invariant Sample Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for analyzing defects in photovoltaic cells fail to retain spatial information and are affected by pseudo-random manipulations, leading to inefficient comparison of defects between samples due to high computational requirements and loss of positional data.
Innovation Solution
A method involving the creation of minimaps with reduced resolution, which account for rotations and symmetries, allowing for effective comparison of defects by determining the distance between minimaps and associating transformations with these distances, thereby preserving spatial information and reducing computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pixel-by-pixel comparison is used to compare defects between samples, then measurement precision is improved, but computation time and computational power requirements increase substantially
Solution Approach 1:
The patent divides the large image into multiple smaller regions or blocks, and performs comparison operations on these segmented regions rather than on the entire image pixel-by-pixel. This segmentation reduces the computational complexity while maintaining the ability to detect defects with sufficient precision.
Solution Approach 2:
The patent introduces intermediate representations such as histograms, feature vectors, or transformed domain representations as mediators between the raw pixel data and the final comparison result. These intermediaries condense the information from large numbers of pixels into compact forms that can be compared efficiently while preserving defect information.
2Measurement precision
If high-resolution images are used to retain spatial information on defects, then measurement precision is improved, but device complexity and computational requirements increase
Solution Approach 1:
The patent transforms the spatial domain image data into another dimension, such as the frequency domain or a compressed feature space, where spatial information is preserved in a more compact form. This dimensionality change allows retention of essential spatial characteristics while reducing the data volume and processing complexity.
Solution Approach 2:
The patent extracts only the essential spatial features and defect characteristics from the full-resolution image, discarding redundant information. This extraction process retains the critical spatial information needed for defect analysis while significantly reducing the complexity of subsequent processing operations.
3Ease of operation
If samples are compared without accounting for rotations and symmetries, then ease of operation is improved, but measurement precision deteriorates due to pseudo-random manipulations
Solution Approach 1:
The patent addresses rotational and symmetry variations by either normalizing the samples to a common orientation or by using comparison methods that are invariant to these transformations. This allows the system to account for pseudo-random manipulations while maintaining operational simplicity.
Solution Approach 2:
The patent changes the parameter space in which comparisons are made, transforming the data into a representation where rotational and symmetry variations are either eliminated or easily accounted for. This parameter transformation enables accurate defect comparison despite variations in sample orientation and manipulation.
Data Source
AI summary
A method of analysis of defects of a type from among a plurality of types of defects between two samples based on an image of each sample characteristic of a type of defect from among the plurality of types of defects includes: for each sample, creating a minimap including bins and representative of a type of defect whose resolution is less than the image of the sample, each bin of the minimap being associated with pixels of the image of the sample and having a score dependent on the pixels and representative of the quantity of a type of defects; determining the distance between each minimap representing the same type or types of defects, the distance between two minimaps being defined as the minimum distance between two minimaps by considering the following transformations: a rotation and/or a symmetry so that each distance between two minimaps is associated with a transformation.