Non-Uniform Downsampling for Panoramic Image Stitching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional downsampling techniques for panoramic images result in loss of image resolution, which affects the quality of image stitching, especially when images need to be scaled for traditional encoders/decoders, and this issue is exacerbated by higher resolution image sensors.
Innovation Solution
A method and apparatus for non-uniform downsampling of captured images, where the processing apparatus receives images from multiple sensors, non-uniformly downsamples them based on the resolution density curve, and encodes the images, preserving more information in areas of interest while allowing for greater flexibility in image processing and stitching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If uniform downsampling is applied to captured panoramic images to fit encoder limitations, then the images can be encoded within standard line size limitations, but image resolution is lost and stitching quality deteriorates
Solution Approach 1:
The patent applies different downsampling factors to different regions of the panoramic image based on a resolution density curve. Regions with higher importance (such as overlapping areas used for stitching) are downsampled less, while less critical regions are downsampled more aggressively. This local differentiation resolves the contradiction by preserving resolution where it matters most for stitching while still achieving overall size reduction for encoder compatibility.
2Measurement precision
If higher resolution image sensors are used to capture panoramic images, then image quality and detail are improved, but the downsampling requirement becomes more severe and resolution loss increases
Solution Approach 1:
The patent changes the downsampling parameter from a uniform fixed value to a variable value that depends on the resolution density curve and regional importance. By dynamically adjusting the downsampling factor based on local image characteristics and stitching requirements, the system preserves more information from high-resolution sensors while still achieving the necessary size reduction for encoding.
3Use of energy by stationary object
If uniform scaling is performed to reduce image size for thermal or battery considerations, then power consumption and device heating are reduced, but image resolution and stitching accuracy deteriorate
Solution Approach 1:
The patent differentiates between critical and non-critical regions when applying downsampling. Overlapping regions that require high stitching accuracy are downsampled with smaller factors, while non-overlapping regions are downsampled more aggressively. This selective approach reduces overall data size and power consumption while maintaining stitching accuracy in critical areas.
Data Source
AI summary
Apparatus and methods for the non-uniform downsampling of captured panoramic images. In one embodiment, a computing device is disclosed that includes a processing apparatus and a non-transitory computer readable apparatus comprising a storage medium have one or more instructions stored thereon. The one or more instructions, when executed by the processing apparatus, being configured to: receive captured images, the captured images obtained using two or more image sensors; non-uniformly downsample the received captured images; and encode the non-uniformly downsampled images. In some variants, the non-uniformly downsampled images take into account a desired area of interest within the captured images. In some implementations, the computing device includes an image capture device. Methods and non-transitory computer readable apparatus are also disclosed.