Road Image Annotation Using 3D Vehicle Path Projection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The manual process of annotating road images for training autonomous vehicle systems is time-consuming and costly, limiting the number of training images that can be generated and affecting the accuracy of road structure detection components.
Innovation Solution
A method for automatically or semi-automatically annotating road images using a system that processes a sequence of images to reconstruct a vehicle's path in 3D space, determining expected road structure, and generating road annotation data through geometric projection, reducing the need for human effort and equipment costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual hand annotation is used to create annotated training images, then annotation accuracy can be maintained, but the time and cost required increases significantly
Solution Approach 1:
The patent uses the vehicle's actual trajectory as a template to automatically generate road structure annotations by copying and adapting the path information across multiple images. This allows accurate annotations to be generated automatically without manual intervention for each image.
Solution Approach 2:
The system performs preliminary processing by capturing the vehicle trajectory once and using it to pre-generate annotations for multiple images. This preliminary action eliminates the need for repeated manual annotation work while maintaining consistency and accuracy across the dataset.
2Manufacturing precision
If manual hand annotation is used to create annotated training images, then annotation quality can be ensured, but the number of images that can be annotated decreases
Solution Approach 1:
The annotation system serves itself by automatically generating road structure annotations using the vehicle's own trajectory data. This self-service approach eliminates the bottleneck of manual annotation while maintaining high quality through the geometric consistency of the trajectory-based method.
Solution Approach 2:
The vehicle trajectory serves multiple functions: it provides both the path information and the reference framework for generating annotations across all images. This multi-functionality allows a single data source to produce consistent, high-quality annotations for the entire dataset.
3Reliability
If manual annotation processes are used, then accurate road structure detection can be achieved, but equipment costs and manual effort increase
Solution Approach 1:
The patent replaces the mechanical process of manual annotation with an automated computational system that uses geometric projection and trajectory analysis. This substitution maintains detection reliability while eliminating the complexity and cost of manual annotation processes.
Data Source
AI summary
A method of annotating road images, the method comprising implementing, at an image processing system, the following steps: receiving a time sequence of two dimensional images as captured by an image capture device of travelling vehicle; processing the images to reconstruct, in three-dimensional space, a path travelled by the vehicle; using the reconstructed vehicle path to determine expected road structure extending along the reconstructed vehicle path; and generating road annotation data for marking at least one of the images with an expected road structure location, by performing a geometric projection of the expected road structure in three-dimensional space onto a two-dimensional plane of that image.