Stereo Depth Maps With Pixel Confidence for Vehicle Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing stereo vision systems lack the ability to provide reliable depth estimates with confidence levels, leading to uncertain decision-making in autonomous vehicles and driver assistance systems, which can result in decreased passenger safety and comfort.
Innovation Solution
A system that combines stereo vision with confidence mapping, using disparity features, prior images, cost curves, and local properties to generate high-resolution depth information and confidence data, enabling reliable sensor fusion and improved decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If stereo vision systems provide depth estimates without confidence levels, then the system complexity is reduced, but the reliability of depth information decreases
Solution Approach 1:
The patent segments the depth information output into two distinct components: depth estimates and confidence levels. Each pixel's depth information is divided into the depth value itself and an associated confidence metric, allowing the system to provide comprehensive information without increasing overall system complexity, as the segmentation occurs at the data output level rather than requiring additional hardware or processing complexity
Solution Approach 2:
The patent introduces confidence levels as an intermediary element between the depth estimation process and the decision-making system. This intermediary provides quantitative information about the reliability of each depth estimate, enabling downstream systems to make more informed decisions without requiring changes to the core stereo vision algorithm
2Loss of information
If stereo vision systems provide depth estimates without confidence levels, then the loss of information is reduced, but the measurement precision decreases
Solution Approach 1:
The patent adds a new dimension to the depth information by introducing confidence levels as a second component alongside depth values. This transforms the output from a single-dimensional depth map to a two-dimensional structure where each pixel contains both depth and confidence, thereby reducing information loss while enhancing measurement precision through the additional reliability metric
3Device complexity
If driver assistance systems use sensor information without confidence data, then the device complexity is reduced, but the reliability of decisions decreases
Solution Approach 1:
The patent performs the confidence assessment action in advance, during the depth estimation process itself. By calculating confidence levels alongside depth values and providing both to the driver assistance system, the reliability information is prepared beforehand, allowing the decision-making system to use this pre-computed confidence data without adding complexity to its own structure
Data Source
AI summary
An automated vehicle assistance system is provided for supervised or unsupervised vehicle movement. The system includes a control system and a first sensor system. The first sensor system may receive first image data of a scene and may output a first disparity map and a first confidence map based on the first image data. The control system may output a video stream based on the first disparity map and the first confidence map. The vehicle assistance system also may include a second sensor system that receives second image data of at least a portion of the scene that outputs a second confidence map based on second image data. The video stream may include super-frames, with each super-frame including a 2D image of the scene, a depth map corresponding to the 2D image, and a certainty map corresponding to the depth map.