Real-Time Roadway Object Detection Using Superpixel Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current object detection systems for vehicles require computationally expensive three-dimensional vision systems, making them commercially infeasible for real-time operation, especially in environments without geo-reference information.
Innovation Solution
A method using a camera sensor and onboard computer to process real-time images by dividing them into superpixels, merging similar ones, generating prior maps, drawing bounding boxes, and performing feature extraction and categorization to identify objects, reducing processing time and cost by using a two-dimensional vision system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If three-dimensional vision systems with large file sizes are used for object detection, then object detection capability is improved, but computational cost and processing time increase significantly
Solution Approach 1:
The patent divides the image processing task into segments: first identifying road regions using prior maps, then extracting features only from those specific regions. This segmentation approach maintains object detection capability while reducing the overall computational burden by avoiding processing of entire three-dimensional images.
Solution Approach 2:
The patent extracts only the necessary features from specific image regions rather than performing feature extraction on entire three-dimensional images. By taking out and processing only relevant road regions identified through prior maps, the system achieves effective object detection with reduced computational cost.
2Measurement precision
If three-dimensional images with large file sizes are used, then object detection accuracy is improved, but processing time increases making real-time operation infeasible
Solution Approach 1:
The patent segments the processing workflow to identify road regions first using prior maps, then performs feature extraction only on those segmented regions. This approach maintains detection accuracy for relevant objects while significantly reducing processing time compared to analyzing entire three-dimensional images.
Solution Approach 2:
The patent performs preliminary actions by generating and using prior maps to identify road regions before performing detailed feature extraction. This preliminary identification step allows the system to focus computational resources on relevant areas, reducing overall processing time while maintaining accuracy.
3Measurement precision
If feature extraction is performed on entire three-dimensional images, then comprehensive object detection is achieved, but computational expense increases making the system commercially infeasible
Solution Approach 1:
The patent extracts and processes only the essential features from identified road regions rather than performing comprehensive feature extraction on entire three-dimensional images. This extraction approach maintains sufficient object detection capability while reducing computational expenses to commercially feasible levels.
Solution Approach 2:
The patent applies partial action by performing feature extraction only on road regions identified through prior maps rather than on entire images. This partial processing approach achieves sufficient object detection for safety applications while reducing computational costs to commercially viable levels.
Data Source
AI summary
The disclosure includes a method that receives a real-time image of a road from a camera sensor communicatively coupled to an onboard computer of a vehicle. The method includes dividing the real-time image into superpixels. The method includes merging the superpixels to form superpixel regions. The method includes generating prior maps from a dataset of road scene images. The method includes drawing a set of bounding boxes where each bounding box surrounds one of the superpixel regions. The method includes comparing the bounding boxes in the set of bounding boxes to a road prior map to identify a road region in the real-time image. The method includes pruning bounding boxes from the set of bounding boxes to reduce the set to remaining bounding boxes. The method may include using a categorization module that identifies the presence of a road scene object in the remaining bounding boxes.