Image Texture Quality Optimization for Computer Vision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image acquisition and processing methods for computer vision are suboptimal, as they often prioritize human vision over computer vision requirements, leading to detrimental effects in applications like autonomous navigation and event detection.
Innovation Solution
The proposed system and method involve analyzing the texture of images captured by imaging devices using processors to adjust parameters such as exposure time and gain, ensuring a desired texture quality that meets specific computer vision application thresholds, using algorithms like FAST or Harris corner detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If image processing algorithms optimize for human vision, then human observation quality is improved, but computer vision application performance deteriorates
Solution Approach 1:
The patent changes the optimization parameters from human vision characteristics to computer vision requirements. Specifically, it adjusts image capture parameters (exposure time, gain, aperture) based on texture quality metrics that are relevant for computer vision algorithms like feature detection and matching, rather than optimizing for human perceptual qualities.
Solution Approach 2:
The patent segments the image quality assessment into different components: human vision quality and computer vision quality (texture quality). It applies different optimization strategies for each segment, allowing independent tuning of parameters to meet specific application requirements without compromising the other aspect.
2Loss of information
If image capture parameters are adjusted to improve texture quality, then computer vision information extraction is improved, but image capture complexity increases
Solution Approach 1:
The patent implements a feedback loop where texture quality metrics are continuously evaluated and used to adjust image capture parameters. The system captures images, analyzes texture quality using feature detection algorithms, and automatically modifies exposure time, gain, or aperture based on the analysis results to maintain optimal texture quality for computer vision applications.
Solution Approach 2:
The system performs self-adjustment of capture parameters based on automatic texture quality assessment. The image capture device autonomously monitors its own output quality and modifies its parameters without external intervention, enabling adaptive optimization for different scene conditions and computer vision application requirements.
3Measurement precision
If exposure time and gain are varied to optimize texture quality, then feature point detection quality is improved, but processing time increases
Solution Approach 1:
The patent establishes predetermined relationships between texture quality metrics and optimal capture parameters. By pre-defining these mappings based on computer vision application requirements, the system can quickly adjust parameters without extensive real-time analysis, reducing processing delays while maintaining feature detection accuracy.
Data Source
AI summary
A system includes a non-transitory computer readable medium storing instructions and a processor coupled to the non-transitory computer readable medium. The processor is configured to execute the instructions to obtain an input indicative of a desired image texture quality, receive an image captured by an image capturing device, analyze texture of the image, and generate a signal to vary or maintain a parameter of the image capturing device based on the analysis of the texture to yield the desired image texture quality.