Driver-Guided Rearview Image Switching to Avoid Camera Occlusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Camera monitor systems (CMS) face issues where subjects captured by multiple cameras with different viewpoints may disappear in composite images due to occlusion, leading to incomplete rearview monitoring for drivers, potentially causing accidents.
Innovation Solution
An image processing apparatus and method that switches and displays images captured by multiple cameras or generates virtual viewpoint composite images based on driver behavior, such as head movements, to ensure continuous and seamless observation of the rearview area.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If a composite image is generated by synthesizing images from multiple cameras with different viewpoints, then a single continuous rearview image can be provided, but subjects may disappear due to occlusion from the reference viewpoint
Solution Approach 1:
The system dynamically switches between different camera viewpoints based on detected driver behavior (head movement direction). When the driver looks left or right, the system transitions from a composite image to a specific camera viewpoint that eliminates occlusion, ensuring subjects remain visible while maintaining image continuity through seamless transitions.
Solution Approach 2:
The rearview monitoring system is segmented into multiple independent camera viewpoints (central rear camera, left rear camera, right rear camera) that can be selectively activated. This segmentation allows the system to choose the optimal viewpoint based on driver behavior, preventing subject disappearance while maintaining continuous monitoring coverage.
2Loss of information
If images are sequentially switched based on driver behavior, then subjects can be continuously observed without occlusion, but the switching may cause visual interruption
Solution Approach 1:
The system performs preliminary detection of driver behavior (head movement) and proactively switches to the appropriate camera viewpoint before the driver's line of sight would be blocked. This advance action ensures continuous subject visibility while maintaining seamless transitions, as the system is already prepared with the correct viewpoint ready for immediate display.
Solution Approach 2:
The system maintains continuous useful action by seamlessly transitioning between different camera viewpoints based on driver behavior. The switching is designed to be smooth and uninterrupted, ensuring that the rearview monitoring function continues without visual interruption or loss of subject visibility throughout the transition process.
3Loss of information
If multiple cameras are installed to provide different viewpoints, then subject visibility is improved, but the system complexity increases
Solution Approach 1:
Multiple cameras are installed with universal functionality to capture rearview images from different viewpoints. Each camera serves multiple purposes: the central rear camera provides the primary composite image, while the left and right rear cameras provide alternative viewpoints for eliminating occlusion. This multi-functionality allows the system to maintain subject visibility without requiring excessive additional hardware.
Data Source
AI summary
Configuration, in which images output to a display unit are switched and displayed in accordance with the behavior of a driver, such as movements of the head of the driver, is achieved. Driver information indicating the behavior of the driver of a moving apparatus and images captured by a plurality of cameras that images a situation around the moving apparatus from different viewpoints are input. The images output to the display unit are switched in accordance with the driver information. The plurality of cameras is, for example, a plurality of rear cameras installed in the rear of the moving apparatus. For example, a direction of the face or line-of-sight of the driver is detected. An image in a direction corresponding to the detected direction of the face or line-of-sight of the driver is selected as an output image, and displayed on the display unit. Alternatively, an image in a direction indicated by a gesture of the driver is selected, and displayed on the display unit.