Vehicle Obstacle Recognition via Edge-Based ROI Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object recognition algorithms require excessive processing power and memory, making them impractical for implementation in embedded systems, particularly in vehicle obstacle recognition systems where real-time image processing is critical.
Innovation Solution
The system focuses on processing only a Region Of Interest (ROI) within the image, converting the image into an edge image, and dividing it into regions to detect areas with high edge component values, thereby reducing the processing load and memory requirements by scanning only the detected ROI.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the object recognition algorithm scans the entire region of the input image, then the object recognition accuracy is improved, but the operation processing amount increases excessively
Solution Approach 1:
The patent divides the input image into multiple regions and performs scanning operations only on selected regions rather than the entire image. This segmentation approach reduces the operation processing amount while maintaining object recognition accuracy by focusing computational resources on areas more likely to contain objects of interest.
Solution Approach 2:
The patent applies different processing strategies to different regions of the image. By identifying and prioritizing certain regions for detailed scanning based on local characteristics, the system achieves high recognition accuracy in critical areas while reducing processing load in less important regions.
2Productivity
If memory is added to handle the excessive operation processing amount, then the object recognition processing capability is improved, but the device complexity increases
Solution Approach 1:
The patent extracts and processes only the necessary portions of the image data rather than handling the entire image. By taking out and focusing on specific regions of interest, the system reduces memory requirements and processing complexity while maintaining effective object recognition capability.
3Reliability
If the scanning operation is performed on the entire image, then the object detection completeness is improved, but the processing time increases
Solution Approach 1:
The patent segments the image into multiple regions and performs scanning operations selectively on these regions. This approach maintains object detection completeness by ensuring all relevant regions are scanned, while reducing processing time by avoiding redundant scanning of entire image areas.
Solution Approach 2:
The patent applies partial scanning action by focusing computational effort on specific regions rather than performing exhaustive scanning across the entire image. This selective approach achieves sufficient detection completeness for practical applications while significantly reducing processing time.
Data Source
AI summary
Provided is an object recognition system. The object recognition system recognizes an object in an ROI of a source image. The object recognition system includes an image change unit and an ROI detection unit. The image change unit receives the source image, and changes the object into an edge image which is represented as an edge line. The ROI detection unit divides the edge image into a plurality of regions, compares a total sum of edge component values of an edge line included in each of the regions and a predetermined threshold value by regions, and detects a region, in which the total sum of edge component values is greater than the threshold value, as the ROI from among the plurality of regions.