Panoramic Image Processing Apparatus with Priority Area Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Generating panoramic images from multiple images captured while panning a camera is challenging due to lens distortion, noise reduction, and image blur issues, particularly when dealing with overlapping images and moving objects.
Innovation Solution
An image processing apparatus that designates priority areas and sets overlapping areas as use-prohibited areas to prevent image blur and reduce distortion effects, allowing for proper panoramic image generation by selectively using pixel values from individual images or averaging them based on predetermined spacing and user input.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the average of pixel values from multiple input images is used at overlapping positions, then noise reduction is improved, but image blur occurs due to lens distortion and moving objects
Solution Approach 1:
The patent divides the image processing into distinct regions: priority areas (where moving objects are detected) and non-priority areas. Different processing rules are applied to each region - single image usage for priority areas to avoid blur, and multi-image averaging for non-priority areas to reduce noise. This segmentation resolves the contradiction by allowing both approaches to coexist in their respective regions.
Solution Approach 2:
The patent applies different quality standards to different local regions of the panoramic image. In priority areas (containing moving objects), only one input image is used to maintain sharpness and avoid blur. In non-priority areas, multiple input images are averaged to reduce noise. This local differentiation allows the system to optimize for noise reduction where appropriate while preventing blur where needed.
2Reliability
If multiple input images are used at overlapping positions, then noise reduction is improved, but distortion effects increase due to lens distortion
Solution Approach 1:
The patent segments the image into priority and non-priority areas based on the detection of moving objects. In priority areas where moving objects are present, only one input image is used to avoid distortion effects. In non-priority areas, multiple images are averaged for noise reduction. This segmentation allows the system to minimize distortion where moving objects are detected while still benefiting from noise reduction in stable regions.
3Manufacturing precision
If single input image is used at overlapping positions, then image blur is prevented, but noise reduction is insufficient
Solution Approach 1:
The patent segments the image processing based on the presence of moving objects in priority areas versus non-priority areas. In priority areas, single image usage prevents blur by avoiding the averaging of different object positions. In non-priority areas, multi-image averaging provides noise reduction. This segmentation resolves the contradiction by applying the appropriate method to the appropriate region.
Solution Approach 2:
The patent implements local quality differentiation where priority areas (with moving objects) use single image processing to maintain sharpness, while non-priority areas use multi-image averaging for noise reduction. This allows the system to prevent blur where moving objects are detected while still achieving noise reduction in stable regions where multiple images are appropriate.
4Ease of operation
If automatic averaging of pixel values is used, then processing simplicity is maintained, but distortion and blur issues arise
Solution Approach 1:
The patent segments the automatic processing into two distinct modes based on region detection: single-image mode for priority areas (preventing blur) and multi-image averaging mode for non-priority areas (noise reduction). This segmentation maintains ease of operation through automated region detection while preventing distortion and blur issues through intelligent selection of processing methods.
Solution Approach 2:
The patent employs feedback through automatic detection of moving objects in input images to determine which regions should use single-image processing versus multi-image averaging. The system continuously monitors the input images, detects moving objects, and adjusts the processing method accordingly. This feedback mechanism maintains processing simplicity while preventing distortion and blur through adaptive decision-making.
Data Source
AI summary
An image processing apparatus includes a first image processor receiving, as a sequence of images, a plurality of images captured by sequentially shifting a shooting position in a predetermined direction, and setting a first overlapping area overlapping a priority area which is to be used preferentially in the sequence of images, as a use-prohibited area with respect to images in the sequence of images other than the first image, the images including the first overlapping area; a second image processor setting a second overlapping area of a second or third image as a use-prohibited area, the second and third images having sequence orders being separated by a predetermined spacing in the sequence of images; and an image generator generating an output image using areas of the plurality of images included in the sequence of images, excluding areas set as the use-prohibited areas by the first and second image processors.