Image Recovery Filter Selection for PSF Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image shooting apparatuses face challenges in maintaining recovery precision when the image shooting distance deviates from the focus location, leading to profile deterioration in recovered images due to varying point spread functions (PSFs), and current methods either require costly distance detection sensors or perform average processing that results in poor precision.
Innovation Solution
An image shooting apparatus and method that utilize pre-calculated recovery filters based on multiple PSFs corresponding to different distances, selecting the optimal PSF for image recovery without deriving the image shooting distance, using a Wiener filter and image sharpness analysis to produce high-precision recovered images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a fixed PSF at the focus location is used for image recovery, then recovery precision at the focus location is maintained, but recovery precision deteriorates when the image shooting distance deviates from the focus location
Solution Approach 1:
The patent segments the continuous PSF variation across different distances into discrete PSF templates corresponding to multiple specific distances. Instead of using a single fixed PSF, the system divides the distance range into segments and prepares recovery filters for each segment, allowing selective application based on the actual shooting distance.
Solution Approach 2:
The patent introduces dynamic selection of PSF templates based on the detected shooting distance. The system dynamically adapts the recovery process by choosing the most appropriate PSF from multiple candidates, making the recovery filter adaptable to varying shooting conditions rather than static.
2Measurement precision
If distance detection sensors or multiple image shooting portions are used to calculate shooting distance for accurate PSF selection, then optimal recovery processing is achieved, but device complexity and cost increase
Solution Approach 1:
The patent enables the imaging system to determine its own shooting distance information by analyzing the captured image data itself, without requiring external distance detection sensors. The system uses image processing techniques to extract depth information directly from the captured images, making the system self-sufficient.
Solution Approach 2:
The patent replaces mechanical distance detection sensors with computational image processing methods. Instead of using physical sensors to measure distance, the system uses algorithmic analysis of image characteristics (such as focus blur analysis) to infer distance information and select appropriate PSF templates.
3Device complexity
If average processing is performed using multiple PSFs to minimize mean square error, then device complexity is reduced, but recovery precision deteriorates compared to using optimal PSF
Solution Approach 1:
The patent performs preliminary preparation of multiple distance-specific PSF templates and recovery filters in advance, organized according to different shooting distances. This pre-computation allows the system to quickly select the most appropriate pre-prepared filter based on detected distance information, avoiding the need for complex real-time optimization calculations.
Data Source
Figure 1
Figure 2A~2B
Figure 3A~3C
AI summary
The invention provides an image shooting apparatus and an image shooting method that obtain a recovered image with high precision by using an optimal point spread function (PSF) without deriving an image shooting distance of a photographed object, an image shooting program, and a recording medium that records the image shooting program. An image shooting apparatus includes an optical system, an image shooting element configured behind the optical system; an image recovery processing portion configured to perform image processing and recovery processing on obtained image data; and a recovered image output portion configured to output a recovered image; and the image recovery processing portion includes a recovery filter storage memory configured to store multiple recovery filters pre-manufactured by using multiple point spread functions; a recovery filter processing portion configured to obtain multiple middle candidate images; and an image comment portion configured to output a middle candidate image with an optimal profile.