Auto-focus prediction using focal surface model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing imaging systems face challenges in quickly acquiring properly focused images, especially when dealing with objects or substrates that have local variations in surface height, as full focus searches are time-consuming and may not account for irregularities, leading to inefficient image capture.
Innovation Solution
The method involves estimating the ideal focal distance at each imaged location using focus data from previous locations, with a processor calculating a representative focal distance based on known focal distances and adjusting for variations, allowing for rapid and accurate auto-focusing without the need for full focus searches at each new location.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full focus search is performed at each location, then focus accuracy is improved, but imaging speed deteriorates
Solution Approach 1:
The system performs a preliminary focus search at selected reference locations to build a focal surface model, then uses this model to predict optimal focal distances at unimaged locations without performing full focus searches there. This preliminary action at strategic points enables rapid imaging elsewhere while maintaining acceptable focus accuracy.
Solution Approach 2:
A focal surface model acts as an intermediary between reference locations with measured focal distances and target locations requiring imaging. The model interpolates and predicts focal distances at unimaged locations based on the spatial relationships established from reference measurements, eliminating the need for time-consuming focus searches at every location.
2Measurement precision
If multiple images are acquired at different focal distances, then focus quality is improved, but acquisition time increases
Solution Approach 1:
The system applies different imaging strategies to different locations: full focus searches are performed only at reference locations where high precision is needed to establish the focal surface model, while unimaged locations use the predictive model for rapid single-image acquisition. This local differentiation optimizes both quality and speed.
Solution Approach 2:
Instead of performing complete focus searches at all locations, the system performs partial focus searches only at strategically selected reference locations. The focal surface model then provides sufficient focus information for the majority of locations, reducing the total number of images needed while maintaining adequate focus quality.
3Manufacturing precision
If focal distance is determined for each location, then image sharpness is improved, but processing complexity increases
Solution Approach 1:
The system replaces the mechanical approach of performing physical focus searches at every location with a computational focal surface model that calculates predicted focal distances based on spatial interpolation. This substitution of computational modeling for repeated mechanical measurement simplifies the overall process while maintaining image sharpness.
Solution Approach 2:
The focal surface model serves multiple functions: it stores focal distance information from reference locations, predicts focal distances for unimaged locations, and provides a unified framework for determining optimal focus settings across the entire imaging area, reducing processing complexity through consolidation.
Data Source
AI summary
The disclosure relates to methods and systems for automatically focusing multiple images of one or more objects on a substrate. The methods include obtaining, by a processor, a representative focal distance for a first location on the substrate based on a set of focal distances at known locations on the substrate. The methods also include acquiring, by an image acquisition device, a set of at least two images of the first location. The images are each acquired using a different focal distance at an offset from the representative focal distance. The methods further include estimating, by a processor, an ideal focal distance corresponding to the first location based on comparing a quality of focus for each of the images, and storing the estimated ideal focal distance and the first location in the set of focal distances at known locations.