Synthetic Aperture Camera for Occluded Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for improving visibility and recognition of partially occluded objects in images, such as those hidden by branches or foliage, require large camera arrays, which are impractical for many applications due to their physical dimensions.
Innovation Solution
A method using a single camera or a small number of cameras that move freely to capture images from different positions, estimating camera movement, and applying synthetic aperture techniques for image registration and fusion, allowing for improved visibility and recognition of occluded objects without the need for a large camera array.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a large camera array is used to improve visibility of occluded objects, then recognition accuracy improves, but device size and complexity increase
Solution Approach 1:
The patent combines multiple images captured from different positions and orientations into a single synthesized image. By merging information from multiple viewpoints, the system achieves comprehensive visibility of occluded objects without requiring simultaneous operation of multiple cameras, thus reducing device complexity while maintaining recognition accuracy.
Solution Approach 2:
The patent employs dynamic image capture where the camera moves to different positions and orientations to capture images of the target area. This dynamic approach allows a single camera to gather information that would traditionally require multiple static cameras, reducing system complexity while improving occlusion handling.
2Reliability
If multiple cameras are positioned at fixed locations to capture occluded objects, then visibility improves, but system portability deteriorates
Solution Approach 1:
The system transitions from static multi-camera setups to a dynamic single-camera approach. The camera moves freely to capture images from multiple positions, making the system portable and adaptable to different environments while maintaining reliable visibility of occluded objects through computational synthesis.
Solution Approach 2:
A single camera performs multiple functions that would traditionally require multiple fixed cameras. By capturing images from various positions and synthesizing them computationally, one camera achieves the multi-view capability of a camera array, enhancing portability without sacrificing visibility reliability.
3Loss of information
If images are captured from multiple positions to reveal occluded objects, then completeness of information improves, but processing complexity increases
Solution Approach 1:
The patent introduces an image synthesis module as an intermediary that processes multiple captured images. This module aligns, fuses, and synthesizes images from different positions into a single comprehensive output, managing processing complexity systematically while ensuring complete information about occluded objects is preserved.
Solution Approach 2:
The system performs preliminary image capture from multiple positions before final synthesis. By gathering all necessary image data first and then processing them together, the system ensures complete information collection while organizing the processing workflow to manage complexity effectively.
Data Source
AI summary
Method for improving the visibly of objects and recognizing objects in a set of images recorded by one or more cameras, the images of said set of images being made from mutual different geometric positions, the method comprising the steps or recording a set or subset of images by means of one camera which is moved rather freely and which makes said images during its movement, thus providing an array of subsequent images, estimating the camera movement between subsequent image recordings, also called ego-motion hereinafter, based on features of those recorded images, registering the camera images using a synthetic aperture method, recognizing said objects.