Image Compositing With Geometric Transformation for Obstacle Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image composition technologies struggle to generate composite images that effectively cancel out obstacle images, especially when imaging conditions significantly change or obstacle characteristics are unclear, and they often require manual operations for correspondence point designation.
Innovation Solution
An image composition apparatus and method that automatically specifies target object regions and obstacle-free regions, uses geometric transformation to generate composite images, and allows for display switching between transformed object and surrounding regions, utilizing both manual and image analysis for correspondence point determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual operation is used to designate correspondence points for image compositing, then the compositing accuracy can be improved, but the operation complexity and time consumption increase
Solution Approach 1:
The system performs automatic correspondence point designation through image analysis, allowing the system to serve itself without requiring manual intervention. The correspondence points are automatically extracted from the images based on feature detection and matching algorithms, eliminating the need for users to manually designate points while maintaining compositing accuracy
Solution Approach 2:
The manual mechanical operation of designating correspondence points is replaced with an automated image analysis system. The system uses computer vision algorithms to automatically identify and match correspondence points between images, substituting the manual mechanical process with an automated computational process that reduces both time and operational complexity
2Productivity
If automatic correspondence point designation is used, then the operation time is reduced, but the compositing precision deteriorates
Solution Approach 1:
The system replaces manual point designation with automated image analysis that uses feature detection and matching algorithms. This automated system quickly identifies correspondence points by analyzing image features such as edges, corners, and distinctive patterns, achieving both high processing speed and maintained accuracy through computational methods
Solution Approach 2:
The system adjusts various parameters in the image analysis process, including feature detection thresholds, matching criteria, and transformation model parameters. By optimizing these parameters, the system achieves accurate correspondence point designation automatically, balancing processing speed with compositing precision without requiring manual intervention
3Manufacturing precision
If multiple images are composited with obstacle cancellation, then the image quality is improved, but the device complexity increases
Solution Approach 1:
The system extracts and removes obstacle regions from the composite image by identifying areas where obstacles appear in some images but not others. The obstacle cancellation function selectively extracts only the necessary image regions from multiple sources, combining them while excluding obstacle areas, thereby improving image quality without requiring complex additional hardware
Solution Approach 2:
The image compositing device performs multiple functions including correspondence point designation, geometric transformation, and obstacle cancellation using a unified system architecture. The same image processing unit handles both the compositing operation and the obstacle detection/removal, reducing overall system complexity while maintaining high image quality through multi-functional integration
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An image composition apparatus includes an image obtaining unit 22 that obtains a first image and a second image obtained by imaging a target object and an obstacle from a first direction and a second direction, a region specifying unit 34 that specifies a target object region by receiving a region designation indicating the target object region in which the target object is captured and the obstacle is not captured on the first image and the second image displayed on a display screen of a display unit 24, a correspondence point information obtaining unit 36 that obtains correspondence point information indicating correspondence points between the first image and the second image, a geometric transformation information obtaining unit 38 that obtains geometric transformation information based on the correspondence point information, a geometric transformation unit 40 that geometrically transforms the target object region of the second image based on the geometric transformation information, and an image composition unit 42 that generates a composite image by compositing the target object region of the first image with the geometrically transformed target object region of the second image.