Image Sensor Readout Control for Depth Map Interpolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image capturing apparatuses face a trade-off between reducing system load and power consumption, which can lead to decreased pixel signal readout and potentially deteriorate object detection accuracy.
Innovation Solution
An image capturing apparatus with a readout unit, a readout control unit, an information acquisition unit, and an interpolation unit that controls the region for reading out image signals with different viewpoints, acquires depth information, and records it along with region information, allowing for interpolation of depth information in thinned regions to maintain accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by stationary object
If the distance information acquisition region is thinned to reduce system load and power consumption, then the pixel signal readout amount decreases, but object detection accuracy deteriorates
Solution Approach 1:
The patent creates a copy of depth information through interpolation. The readout control unit reads out image signals only from a thinned acquisition region, but the interpolation unit generates a complete depth map by copying and extrapolating the read information to fill in the thinned regions, thus maintaining full-resolution object detection without reading all pixel signals.
Solution Approach 2:
The patent extracts only the essential information needed for depth detection from the image sensor. By thinning the acquisition region to specific pixel rows or columns, the system extracts minimal necessary signal data while the interpolation unit reconstructs the complete depth information, reducing power consumption while maintaining detection accuracy.
2Productivity
If the distance information acquisition region is thinned to reduce system load, then the pixel signal readout amount decreases, but object detection accuracy deteriorates
Solution Approach 1:
The interpolation unit creates a copy of the depth information from the thinned acquisition region and extrapolates it to fill the entire image area. This allows the system to process only a subset of pixel data while maintaining full-resolution depth maps for object detection, thereby increasing processing speed without sacrificing accuracy.
Solution Approach 2:
The interpolation unit performs preliminary processing by pre-calculating depth values at the thinned acquisition regions and using these as seeds for generating the complete depth map. This preliminary action reduces the amount of real-time processing needed, improving overall processing speed while maintaining detection precision.
3Measurement precision
If all image signals are read out from the image sensor, then object detection accuracy is maintained, but the data amount increases and processing load increases
Solution Approach 1:
The readout control unit extracts only the necessary image signals from specific pixel rows or columns in the acquisition region, rather than reading all signals from the entire sensor array. This extraction approach maintains sufficient depth information for accurate object detection while significantly reducing the data volume and processing load.
Solution Approach 2:
The patent segments the image sensor into distinct functional regions: the distance information acquisition region with thinned pixel selection and the regular image capture region. This segmentation allows the system to process depth information and color information separately, reducing overall processing complexity while maintaining detection accuracy.
Data Source
AI summary
An image capturing apparatus having an image sensor, that reads out an image signal from pixels of the image sensor, controls a region in which a plurality of image signals having different viewpoints are read out from pixels of the image sensor, acquires depth information using the image signals read out from the region, and records image information in which the image signals, the depth information and information regarding the region are associated with each other.