Image Processing Using Dual-Light Subject Detection for Fog-Haze Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques struggle to enhance visibility of subjects obscured by fog or haze, particularly when conventional subject detection methods fail to accurately identify main subjects in such conditions.
Innovation Solution
An image processing apparatus that utilizes both visible and invisible light cameras to detect subjects, integrating evaluation values from both sources to determine a main subject region, and applies gradation correction based on these evaluations to improve visibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If correction processing is performed on the entire image using conventional luminance correction curves, then the overall visibility of the image is improved, but the correction is not suitable for the main subject and may even degrade subject visibility
Solution Approach 1:
The image is segmented into multiple regions: main subject region, foreground region, and background region. Different gradation correction curves are applied to each region based on their respective characteristics and fog/haze conditions, allowing subject-specific optimization while maintaining overall image quality
Solution Approach 2:
Different gradation correction characteristics are applied to different regions of the image. The main subject region receives correction optimized for subject visibility, while foreground and background regions receive corrections tailored to their specific atmospheric conditions and visual requirements
2Manufacturing precision
If known subject detection techniques are used to identify the main subject, then region-specific correction can be applied, but detection itself cannot be performed if the subject is strongly covered in fog or haze
Solution Approach 1:
A fog/haze detection unit is introduced as an intermediary that operates before subject detection. It detects the presence and extent of fog/haze in different image regions, providing information that enables robust subject detection even when subjects are obscured, by allowing the system to adjust detection parameters or use alternative detection strategies based on atmospheric conditions
Solution Approach 2:
Fog and haze detection is performed preliminarily before subject detection and gradation correction. This preliminary detection of atmospheric conditions allows the system to prepare appropriate detection and correction strategies, ensuring reliable subject identification even when subjects are partially obscured by fog or haze
3Productivity
If a single gradation correction curve is applied to the whole image, then the processing is simple and fast, but the correction is not optimized for the main subject and reduces image clarity
Solution Approach 1:
The image processing is segmented into region identification, fog/haze detection for each region, and application of different gradation correction curves. While more complex than single-curve correction, this segmentation enables subject-optimized correction that significantly improves subject visibility and overall image quality
Solution Approach 2:
The system applies gradation correction selectively to different regions rather than uniformly to the entire image. By focusing correction efforts on the main subject region with optimized parameters while applying appropriate corrections to foreground and background regions, the system achieves superior subject visibility enhancement
Data Source
AI summary
An image processing apparatus comprises a first detection unit configured to detect a subject from a first image obtained by image capturing using visible light, a second detection unit configured to detect a subject from a second image obtained by image capturing using invisible light, and a correction unit configured to perform gradation correction on the first image based on a detection result of the subject by the first detection unit and a detection result of the subject by the second detection unit.