Route Surrounding Image Synthesis for Distortion-Free Side Views

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing driving assistance systems distort or stretch three-dimensional objects in overhead or bird's-eye view images due to viewpoint conversion without considering their original shape, especially when capturing images from multiple angles.

Innovation Solution

A driving assistance device that cumulatively stores and synthesizes only a part of the captured images from side cameras, extracting lateral regions of the vehicle, to generate a route surrounding image without distortion, using vehicle information for accurate display.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If viewpoint conversion is performed on captured images to generate overhead or bird's-eye view images, then visibility of the surrounding environment is improved, but three-dimensional objects are displayed distorted or stretched

Engineering Contradiction:
Improvevisibility of surrounding environmentVSAvoidshape accuracy of three-dimensional objects
Core Design Contradiction:
Loss of informationVSManufacturing precision

Solution Approach 1:

The patent segments the captured image into multiple regions based on distance from the vehicle, applying different processing methods to each region. Specifically, it divides the image into a first region (closer to vehicle), second region (intermediate distance), and third region (farther distance), and selectively extracts and synthesizes images from different cameras for each region to maintain shape accuracy while providing overhead view functionality.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different image processing qualities to different spatial regions. For regions where shape accuracy is critical (closer to vehicle), it uses extracted images with minimal distortion. For regions where overall coverage is more important (farther away), it uses synthesized overhead view images. This local differentiation resolves the contradiction between visibility and shape accuracy.

Inventive Principle:
Principle #3Local quality

2Area of stationary object

If all captured images are synthesized for overhead view, then complete surrounding coverage is achieved, but processing complexity and computational load increase

Engineering Contradiction:
Improvecoverage area of surrounding environmentVSAvoidimage processing complexity
Core Design Contradiction:
Area of stationary objectVSDevice complexity

Solution Approach 1:

Instead of synthesizing all captured images uniformly, the patent applies partial synthesis only where necessary. It extracts specific images from specific cameras for specific regions, performing synthesis only for the third region (farther distances) where overhead view provides added value, while leaving closer regions as direct extracted images. This reduces processing complexity while maintaining adequate coverage.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The patent segments the image processing task by region, applying different synthesis strategies to different spatial zones. This segmentation allows the system to achieve complete surrounding coverage through selective synthesis rather than processing all images uniformly, thereby reducing overall computational complexity.

Inventive Principle:
Principle #1Segmentation

3Area of moving object

If wide-angle lens is used to eliminate blind spots, then field of view is expanded, but image distortion increases

Engineering Contradiction:
Improvefield of view of imaging deviceVSAvoidimage distortion
Core Design Contradiction:
Area of moving objectVSManufacturing precision

Solution Approach 1:

The patent extracts specific regions and specific images from the captured wide-angle images, removing the distorted portions. It identifies and extracts undistorted or less-distorted regions (such as central regions or specific angular zones) and uses these extracted portions for critical displays, effectively taking out the useful information while leaving behind the distorted areas.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent changes the parameter of image selection based on spatial parameters. It selects different images from different cameras depending on the region being displayed, effectively changing the optical parameters (which camera is used) to minimize distortion for each specific display region while maintaining overall wide coverage.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4624258A1Driving assistance device
Publication Date: 2025.10.01 AISIN CORP
  • EP4624258A1 patent drawingFigure 1
  • EP4624258A1 patent drawingFigure 2
  • EP4624258A1 patent drawingFigure 3

AI summary

A driving assistance device (1) includes: an image storage unit (31) configured to cumulatively store captured images, each obtained by capturing an image of a surrounding environment by an imaging device (6, 7, 8A, 8B) provided in a vehicle (2), in association with vehicle information when capturing the respective captured images; an image synthesizing unit (31) configured to synthesize the cumulatively stored captured images based on the vehicle information to generate a route surrounding image (44) indicating a surrounding environment along a route on which the vehicle travels; and an image display unit (4) configured to display the route surrounding image along a route on which the vehicle is currently traveling. The imaging device is provided to capture an image of a region including a side region located laterally with respect to the vehicle. The image storage unit extracts and stores only a part, which includes the side region, in the captured image captured by the imaging device.