Vehicle Surround-View Blind Spot Filling for Natural Image Rendering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Surround-view monitoring systems in vehicles often have blind spots due to camera installation positions, leading to areas with no image data, which are typically masked in black or with gradation, making the images appear unnatural.
Innovation Solution
A method and apparatus that utilize images from the surroundings of the blind spot to generate a corrected surround-view image by setting regions of interest, cropping and resizing virtual images, and synthesizing them with the vehicle image to replace the blind area, applying filters and adjusting brightness for a natural appearance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the blind area is masked in black or with gradation, then the blind spot is clearly indicated, but the SVM image appears unnatural and quality deteriorates
Solution Approach 1:
The patent creates a virtual image by copying and transforming image data from regions surrounding the blind spot. The virtual image is generated by cropping, resizing, and rotating the surrounding region to fill the blind area, thereby replicating visual information that would naturally be present rather than masking the area.
Solution Approach 2:
The patent introduces a virtual image as an intermediary element between the camera's actual field of view and the final SVM image. This virtual image acts as a mediator that fills the blind spot area with synthesized visual data, bridging the gap between captured images and the desired complete surround view.
2Area of stationary object
If multiple cameras are used to cover more area, then the blind spot coverage improves, but the device complexity and cost increase
Solution Approach 1:
The patent transitions from a two-dimensional camera array to a three-dimensional spatial solution by utilizing depth information and spatial relationships. The virtual image is generated by mathematically transforming coordinates from surrounding regions, effectively adding a dimensional transformation approach rather than simply adding more cameras.
Solution Approach 2:
The patent replaces the mechanical approach of adding more physical cameras with a computational image processing solution. Instead of mechanically increasing camera quantity, the system uses algorithms to synthesize missing visual information from existing camera data, substituting computational complexity for hardware complexity.
3Device complexity
If the blind area is left uncorrected, then the processing complexity is reduced, but the image clarity and visibility deteriorate
Solution Approach 1:
The patent performs preliminary actions by pre-processing the surrounding region data before final image composition. The virtual image is generated in advance through cropping, resizing, and rotation operations, allowing the blind spot to be filled seamlessly during the final image assembly without adding significant complexity to the main processing pipeline.
Data Source
AI summary
An apparatus for surround-view monitoring for a vehicle includes a plurality of cameras, a display, and a controller. The controller is configured to generate a surround-view image containing a blind area by synthesizing a plurality of images obtained from the plurality of cameras. The controller is also configured to correct the blind area by utilizing image data of the surroundings of the blind area in the generated surround-view image. The controller is further configured to output the corrected surround-view image to the display.