Focus Detection via 2D-to-1D Image Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing focus detection systems face challenges in accurately detecting focus under diminished brightness conditions, often resulting in inappropriate computation results when the same subject does not appear at each pixel or when images approximated by lines oblique to the correlation direction are encountered.

Innovation Solution

The proposed solution involves an image capturing apparatus with a two-dimensional array of focus detection sensors and correlation calculation means that calculate the correlation between two-dimensional images shifted in the pupil dividing direction, while also considering the direction perpendicular to it, to obtain a single focus measurement value effectively.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Illumination intensity

If image signals are synthesized by combining multiple pairs of image signals in the direction perpendicular to the pupil dividing direction, then focus detection capability under diminished brightness is improved, but inappropriate computation results occur when the same subject does not appear at each pixel or when images approximated by lines oblique to the correlation direction appear

Engineering Contradiction:
Improvefocus detection capability under diminished brightnessVSAvoidfocus measurement accuracy
Core Design Contradiction:
Illumination intensityVSMeasurement precision

Solution Approach 1:

The patent segments the two-dimensional images into multiple one-dimensional image signals by dividing along the pupil dividing direction. This segmentation allows the system to process images that cannot be properly handled by conventional two-dimensional correlation methods, particularly when subjects are at varying distances or when lines oblique to the correlation direction are present. By converting 2D images to 1D signals through strategic segmentation, the system achieves reliable focus measurement in complex scenarios while maintaining low-light detection capability.

Inventive Principle:
Principle #1Segmentation

2Illumination intensity

If hardware-based improvements are made to raise sensor sensitivity or lengthen signal accumulation time, then focus detection under diminished brightness is improved, but device complexity and cost increase

Engineering Contradiction:
Improvefocus detection capability under diminished brightnessVSAvoidhardware complexity
Core Design Contradiction:
Illumination intensityVSDevice complexity

Solution Approach 1:

The patent replaces hardware-based solutions (increased sensor sensitivity, longer accumulation time, divided-pupil arrangements) with a software-based image processing method. By using correlation calculation on segmented one-dimensional image signals, the system achieves focus detection under diminished brightness without modifying the physical sensor or optical system, thereby avoiding increased device complexity and cost.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Ease of manufacture

If conventional correlation computation is performed on two-dimensional images, then processing is simple, but focus detection accuracy deteriorates when subjects are at long and short distances or under diminished brightness

Engineering Contradiction:
Improveprocessing simplicityVSAvoidfocus detection accuracy
Core Design Contradiction:
Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent transforms the problem from two-dimensional image correlation to one-dimensional signal correlation by segmenting the 2D images along the pupil dividing direction. This dimensionality reduction simplifies the computation while simultaneously improving accuracy for subjects at varying distances and under low-light conditions. The 1D correlation approach is both simpler to implement and more robust than 2D correlation for the specific problem of focus detection.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach enhances focus detection accuracy under low-light conditions by improving the signal-to-noise ratio and appropriately selecting the focus-measurement target, even in environments with subjects at varying distances.

Implementation Method 1

focus detection sensors arranged two-dimensionally for photoelectrically converting images of the pupil dividing means

Methodology Applied
Scientific EffectPhotoelectric conversion: Photoelectric Effect

Data Source

PatentUS8634015B2Image capturing apparatus and method and program for controlling same
Publication Date: 2014.01.21 CANON KK
  • US8634015B2 patent drawing
  • US8634015B2 patent drawing
  • US8634015B2 patent drawing

AI summary

An image capturing apparatus includes an image sensor provided in the neighborhood of an image plane of a photographing optical system for forming the image of a subject; pupil dividing member for forming images of a pair of light beams that have passed through different pupil regions of the photographing optical system; focus detection sensors arranged two-dimensionally for photoelectrically converting images of the pupil dividing member; and a correlation calculation unit which, in order to obtain one focus measurement value from one pair of two-dimensional images generated by the focus detection sensors and extending in the pupil dividing direction of the pupil dividing member and in a direction perpendicular to the pupil dividing direction, calculates the correlation between the pair of two-dimensional images while shifting the pair of two-dimensional images in the pupil dividing direction.