Defocus Map Alignment for Camera Shake Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing techniques for generating defocus maps from images with different exposure times are affected by camera shaking and increased noise, leading to reduced accuracy, especially when luminance differences are significant.
Innovation Solution
An image processing apparatus that acquires and aligns first and second sets of image data with different shooting times, generates spatial distribution maps of defocus amounts, and performs composing processing to enhance the accuracy of defocus maps, while correcting for positional deviations and noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple images with different exposure times are used to generate defocus maps, then the accuracy of defocus map is improved, but the impact of camera shaking and noise increases
Solution Approach 1:
The patent applies preliminary action by performing alignment processing on the first and second map data before composing them. The alignment unit acquires the map data from the generating unit and performs alignment to correct positional deviations caused by camera shaking. This preliminary alignment ensures that when the map data are composed, they are properly registered, thereby maintaining defocus map accuracy while mitigating the negative effects of camera shaking.
Solution Approach 2:
The patent introduces an alignment unit as an intermediary between the map data generation and composition processes. This intermediary component processes the map data to correct positional deviations before the composing operation, acting as a mediator that resolves the conflict between utilizing multiple images for improved accuracy and dealing with camera shaking-induced misalignment.
2Measurement precision
If images with different exposure times are composed, then luminance difference accuracy is improved, but noise amount increases
Solution Approach 1:
The alignment unit serves as an intermediary that processes map data before composition, ensuring proper alignment and reducing noise propagation. By performing alignment on the generated map data rather than directly on the raw images, the system mitigates noise impact while maintaining the ability to accurately represent luminance differences in the final defocus map.
Solution Approach 2:
The patent replaces direct image composition with a two-stage process: first generating defocus maps from images with different exposures, then aligning and composing these maps. This substitution of the mechanical composition process with a more sophisticated map-based approach reduces noise impact while preserving luminance difference accuracy.
Data Source
AI summary
An image processing unit of an image pickup apparatus acquires a plurality of sets of image data that have different shooting times. A defocus map generating unit generates each of a plurality of defocus maps that represent a spatial distribution of a defocus amount by using two or more sets of image data. A positional deviation correction unit performs alignment of a plurality of defocus maps through the geometrical deformation processing by using a control parameter that has been acquired from a positional deviation correction control parameter calculation unit. A composing unit composes the plurality of defocus maps that have undergone alignment. A defocus map shaping unit performs the shaping processing on the defocus map after the composing processing, and outputs defocus map data that smoothly changes along the outline of an object.