Panoramic Image Stitching Using Sharpness and Overlap Measures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image capture devices, such as cameras, often introduce blur in panoramic images due to involuntary oscillations of the lens and hand shake, which can result in parts of the panoramic image being out of focus, especially when the shake frequency coincides with the lens and spring assembly's resonance frequency.
Innovation Solution
A method and apparatus for generating panoramic images by selecting and stitching sharp image frame portions from consecutive image frames based on sharpness and overlap measures, rather than relying on overlap alone, to avoid blur and improve image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If image frames are captured during camera movement to create panoramic images, then the field of view and coverage are improved, but image quality deteriorates due to blur from hand shake and lens oscillation
Solution Approach 1:
The panoramic image is segmented into multiple individual image frames captured during camera movement. Each frame is evaluated independently for sharpness and overlap quality, allowing selective use of only the high-quality segments in the final stitched panorama, thus maintaining both wide coverage and high image quality
Solution Approach 2:
The system performs preliminary evaluation of each captured image frame for sharpness and overlap quality before final stitching. This preliminary action identifies and flags blurry or low-quality frames, enabling the selection process to exclude them from the final panoramic image, thereby preventing quality deterioration
2Area of stationary object
If multiple consecutive image frames are stitched to create panoramic images, then the coverage area is improved, but image quality deteriorates due to inclusion of blurry sections from oscillation and hand shake
Solution Approach 1:
The system applies local quality assessment to each individual image frame by evaluating sharpness and overlap measures independently. This allows different portions of the panoramic image to be constructed from frames with locally optimal quality, ensuring that only high-quality sections are included while maintaining comprehensive coverage
Solution Approach 2:
Before stitching, the system performs preliminary evaluation of each frame's sharpness and overlap quality. This preliminary action enables the selection process to identify and exclude blurry sections, ensuring that only high-quality frames contribute to the final panoramic image
3Device complexity
If all consecutive image frames are used for stitching, then the stitching process is simplified, but image quality deteriorates due to resonance frequency oscillation causing out of focus regions
Solution Approach 1:
The system performs preliminary evaluation of each image frame for sharpness and overlap quality before the stitching process. This preliminary action identifies frames affected by resonance frequency oscillation and marks them for exclusion, allowing the stitching process to proceed with only high-quality frames while maintaining relative simplicity
Solution Approach 2:
The system enables the image frames to self-evaluate their own quality through automated sharpness and overlap measurements. Each frame essentially selects itself for inclusion or exclusion based on its own quality metrics, reducing the complexity of external quality assessment while ensuring high image quality in the final panorama
Data Source
AI summary
A method and an apparatus for generating panoramic images are presented. The method includes receiving a set of consecutive image frames. A portion of a first image frame is extracted to configure an intermediate panoramic image. The intermediate panoramic image is augmented by stitching image frame portions to the intermediate panoramic image. The image frame portions are selected from individual image frames in a subset of image frames including consecutive image frames from a second image frame to a penultimate image frame. Each image frame portion is selected based on a sharpness measure and an overlap measure associated with the respective individual image frame. The overlap measure is determined based on a comparison of the respective individual image frame with a current state of the intermediate panoramic image. A portion of a last image frame is stitched to the augmented intermediate panoramic image to configure the panoramic image.