Depth Detection Apparatus Spatial Frequency Confidence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing depth detection methods in digital cameras face challenges in accurately determining confidence for depth calculations due to the influence of spatial frequency, leading to potential errors in depth measurement, especially when spatial frequency distributions are not adequately considered.
Innovation Solution
A depth detection apparatus that generates third and fourth images from first and second images using filters, acquires representative values of spatial frequencies, and determines confidence based on these values, enhancing the accuracy of depth information detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If template matching is used to obtain image shift amount, then depth measurement can be achieved at high speed, but calculation error increases when spatial frequency is low
Solution Approach 1:
The patent changes the parameter used for confidence determination from contrast change amount alone to a composite parameter that includes both contrast change amount and spatial frequency information. This is achieved by calculating spatial frequency distribution in the image and combining it with contrast change amount to determine confidence, thereby resolving the contradiction between speed and precision in depth measurement
2Loss of information
If Fourier transform is used to calculate spatial frequency distribution, then spatial frequency information can be obtained, but calculation load and time period increase
Solution Approach 1:
The patent extracts only the necessary spatial frequency information needed for confidence determination without performing a complete Fourier transform. By selectively calculating spatial frequency distribution only in relevant regions and for specific purposes, the patent reduces calculation load while maintaining the essential spatial frequency information needed for accurate depth measurement
3Device complexity
If confidence is determined based only on contrast change amount, then calculation is simple, but accuracy is insufficient due to ignoring spatial frequency influence
Solution Approach 1:
The patent creates a composite confidence metric by combining contrast change amount and spatial frequency information. This composite approach integrates multiple types of image information to determine confidence, achieving higher accuracy in depth calculation while managing complexity through efficient calculation methods
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 allows for high-accuracy confidence determination in depth detection, reducing calculation errors and improving the speed of the process without the need for complex Fourier transforms, thus providing reliable depth measurements.
Implementation Method 1
pixels each having a depth measuring function (hereinafter also referred to as 'depth measuring pixels') are disposed in part or all of an image sensing device
Data Source
AI summary
Provided is a depth detection apparatus for detecting depth information from a first and a second image based on a light flux passing through different pupil regions, the apparatus including an image generator that generates a third and a fourth image by subjecting the first and the second image to a filter, a depth detector that detects depth information from a relative positional shift amount between the third and the fourth image, a representative value acquirer that acquires a representative value of spatial frequencies of the third or the fourth image from a frequency characteristic of the filter, and a confidence determiner that determines confidence for the depth information from the representative value. The confidence is determined as a value representing higher confidence as the representative value is greater.